Enhancing tunnel sprayer performance through

Full text

Turn on search term navigation

Introduction

Chemical spraying is widely used for crop protection, owing to its higher degree of protection from insect pests and easy handling^1,2. Efficacy of spray depends on spray content reach at target pest. Therefore, various researchers tried to quantify the useful portion of the applied pesticide. In 1986, researcher³ revealed that only 0.003% of the applied pesticide was required to control target caterpillars in collards. However, less than 0.1% of applied pesticides reach the target pest and rest of the pesticides are wasted, and led to degradation of the environment as reported by researcher⁴. To improve pesticide reachability, spraying systems were tested with air assistance mechanism^{5, 6, 7–8}but it can create the problem of off-target losses to environment^{9, 10–11}. For crop protection along with minimum environment degradation, several spraying technologies had been developed, such as air blast sprayers,, electrostatic spraying systems, sensor based sprayer, computer vision based spraying system using drone or satellite mapping, tunnel spraying systems etc. Air blast sprayer facilitates high spray coverage at greater height tree, but cause canopy destruction due to high pressure and high ground loss. The electrostatic sprayer charges the spray droplets and directs it toward the plant leaves surface and reduced chemical use, but high system cost and potential human hazards are sprayer’s limitation due to high voltage, electrical discharge issues and inhalation hazard^{12, 13, 14, 15, 16–17}. On other hand, sensor-based sprayers are very efficient for effective delivery of the pesticide over the canopy and save significant pesticide volume with sustainable environment use, but limited canopy coverage and unstable spray are potential limitation of the sensor-based sprayer^{18, 19, 20, 21–22}. The computer vision-based spraying technology offers precise weed detection and targeted spraying, resulting in less chemical waste and improved applicator safety^{23, 24, 25–26}. It has several significant problems, including synchronization of field input data with release of pesticide, high starting costs, and technological complexity. However, the tunnel sprayer is a distinct recycling system, provided complete canopy coverage and captures surplus pesticide to be not retained by the canopy and saves a considerable fraction of the delivered spray mixture as reported by several researchers^{27, 28, 29, 30, 31, 32–33}. So far, the different aspects of tunnel sprayer had been studied for its efficient working in terms of recycling rate, spray deposits or deposition, spray coverage, drift loss, spray deposition uniformity, and design. Initially, tunnel sprayers were developed without air assistance and reported problems of deficient spray coverage with minimum deep penetration in the canopy^34,35. Thus, these researchers suggested air assistance for uniform spray deposits on both leaf surfaces, lead to the incorporation of different fan into the shield to produce sufficient air flow rate. Several researchers tested the tunnel sprayer with different aspects of air outlets to determine efficient air outlet orientation^30,31,36,37 and air outlets number^{38, 39–40}. To improve potential recovery ability of the tunnel sprayer, air flow trajectory inside the tunnel shield was traced by the researchers using different approaches like modeling approach⁴¹. The design of the tunnel sprayer was not compact, and robust. In 2005, researcher⁴² had employed CFD simulation approach to make more effective for recycling rate, and reduce weight of the system. For quantifying effective air assistance and avoiding unnecessary air assistance in the sprayer, different aspects of air assistance was incorporated in the tunnel sprayer and tested with different air volumetric flow rate⁴³and airflow rate⁴⁴. In addition, air assistance with a closed-loop recycling system^45,46 was adopted in the tunnel sprayers to enhance recovery rate and to prevent environment degradation by reducing drift loss. Since the closed loop system used an air flow input to be contained certain pesticide fraction, which was released by opposite fan, led to improve spray deposits and recycling rate by throwing same air flow volume again and again over the canopy. Off-course, it improves recovery rate at some extent, but reduced the life of fan to be contaminated by the applied pesticide. This contamination fan problem was resolved by adopting external fans with separator screens in the recycling system⁴⁷.

Recent test studies have assessed the performance of tunnel sprayers with different morphological developments. In 2017, the transversal movement suitability of tunnel sprayers as improved within trained vineyards by reducing machine size⁴⁸. The effect of canopy development on spray deposition⁴⁹ was assessed within the tunnel spraying system. The guava morphology was studied and suggested the use of research findings for the design and development of tunnel sprayers^50,51.

The performance of the tunnel sprayer depends on effectiveness of the recycling system, consisted of different constructional components, such as the nozzle, nozzle characteristics, air duct, fins, and tunnel opening which could play a significant role in delivering higher spray droplets, improving spray uniformity, and better recycling rate. Since nozzle characteristics with tunnel opening influences the effective delivery of applied pesticide and deep canopy penetration. The fin impacts the expose area of the each tunnel shield, which may influence the recycling rate by capturing surplus pesticide.

Therefore, this attempt is made to investigate the effect of constructional parameters (nozzle spacing, nozzle angle, tunnel opening, and fin pitch) on droplet deposit, and recycling rate. To quantify exact inputs, numerical optimization was carried out using RSM approach for better recovery rate without sacrificing effective insect pest’s management by higher spray deposits and uniform spatial spray distribution. However, integration of particle swarm optimization (PSO) with artificial neural networks (ANNs) was also performed for identification of optimal constructional parameters against each response. This approach leverages the strengths of both methodologies, allowing for enhanced predictive accuracy and improved decision-making in complex scenarios. By systematically adjusting the constructional parameters, the model can better adapt to varying conditions and yield superior performance outcomes. Moreover, despite the availability of various pesticide spraying technologies, significant gaps remain in achieving both high spray efficiency and minimal environmental impact. Existing tunnel sprayers, although effective in recycling surplus pesticide, often suffer from suboptimal constructional designs, limiting spray coverage, uniformity, and recovery rates. Literature lacks a comprehensive investigation into the synergistic effects of key constructional parameters—such as nozzle spacing, nozzle angle, tunnel opening, and fin pitch—on spray performance. This study aims to bridge that gap by optimizing these parameters using RSM and PSO-ANN models, thus enhancing tunnel sprayer efficiency while addressing limitations in spray deposition, recycling, and environmental sustainability. The novelty of this manuscript lies in the integration of Response Surface Methodology (RSM) and Particle Swarm Optimization (PSO) with Artificial Neural Networks (ANNs) to optimize key constructional parameters of tunnel sprayers, such as nozzle spacing, angle, tunnel opening, and fin pitch, for improved spray efficiency, recycling rate, and environmental sustainability.

Materials and methods

Prototype of tunnel sprayer

A single-row tunnel sprayer was developed with two symmetrical tunnel shields. Each shield [Height×Length×Width (H×L×W): 1250 × 1200 × 200 mm] was fabricated using a MS sheet. For free movement of tunnel shields on the ground surface, both shields were mounted over a trailer unit (Fig. 1), which consisted of four plastic wheels and a trailer frame. These wheels [Diameter (D): 80 mm] were attached at the bottom and each corner of the trailer unit. However, an air duct (L×D: 1200 × 100 mm) was installed vertically and 100 mm away from one side of each tunnel to keep free space for nozzle angle and nozzle spacing adjustment. The bottom end of the air duct (Fig. 1) was closed with a suitable end cap, while its other end is connected to a hose pipe. It has two square openings (60 × 60 mm) and an air direction control mechanism, fitted over these openings to provide suitable air direction according to nozzle angle.

Fig. 1 [Images not available. See PDF.]

Details of tunnel shield mounted with other accessories on test rig set up (1: blower integrated with engine, 2: hose pipe, 3: vertical air duct, 4: square opening in air duct, 5: air regulator valve, 6: fins, 7: hollow cone nozzle, 8: pesticide supply line, 9: trailer unit, 10: tunnel shield, 11: hybrid tree, 12: recycling output pipe, 13: plastic wheel).

The orientation of these openings was kept as per recommendations by various researchers^31,36,52. One hollow cone nozzle (Fig. 1) was placed in front of each square opening of the air duct using the nozzle angle adjustment mechanism, so that maximum delivery of spray droplets over the canopy can be possible with the same air volume. A number of acrylic fins (L×W×T: 1100 × 150 × 4 mm, each) were installed in standing position and parallel to the tunnel shield’s height using a fin pitch adjustment mechanism in order to capture spray droplets (Fig. 1). A recycling basin of each shield was used for the collection of recycling spray mixture. Both tunnel shields were interconnected from top using a tunnel opening adjustment mechanism, and consisted of two frames, namely the shield mounting frame and the blower mounting frame. The shield mounting frame was fitted with the shields, while the blower mounting frame was fixed with a test trolley set up and shields as well. A petrol engine (3 kW, Makita, Japan) integrated with a centrifugal blower was mounted over the blower mounting frame and connected to the tunnel shield (Fig. 1). The blower outlet was bifurcated in the Y-section to supply sufficient air quantity to the air duct via a hose pipe. In this study, air velocity was measured using a digital propeller anemometer (KUSAM-MECO, India) at the center of both shields and in front of square openings. Average air velocity (15 m/s) was maintained in the whole experiment using an air regulator valve, i.e., fixed 450 mm below the blower outlet.

Two ceramic ConeJet hollow cone nozzles (Model: TXA8002VK, TeeJet, USA) were fitted with each tunnel shield, and recorded 1.35 l/min liquid flow rate per nozzle with constant (0.98 MPa) operating pressure. The piston pump was powered by an induction motor (5.6 kW), and lifted the desired volume of spray mixture from the main tank having 800-liters storage capacity, and delivered sufficient volume to these nozzles with surplus liquid returning arrangement to the main tank.

The line diagram of the chain and sprocket arrangement on a test trolley set up was shown in Fig. 2, and allowed to tunnel sprayer for backward and forward movement over both hybrid trees during each experimental run. It converted the rotary motion of the electric motor to the linear motion of the tunnel sprayer over a test trolley set up through a belt and pulley arrangement. The speed of the tunnel spraying operation was kept constant (0.97 m/s) throughout the experiment using a flux vector drive, which permitted the regulation of motor rotational speed.

Fig. 2 [Images not available. See PDF.]

Line diagram speed control unit on test trolley set up.

Two hybrid trees were fabricated on the basis of the morphological characteristics of natural guava trees as per the methodology used by researchers^50,51where leaf area index were estimated in the field condition by considering ground surface area and total leaf area i.e., calculated by multiplication of average leaf area (based on 60 observations) and number of leaves. To achieve LAI 1.6, the number of leaves was estimated on the basis of the LAI for 1.6 divided by the multiplication of the ground surface area (as per field condition, i.e., plant-to-plant distance of 2 m and assuming the same row-to-row spacing in the laboratory to create simulated field condition) and the average leaf area of a single leaf. It means that this study is conducted in the laboratory conditions, but carried out over the hybrid tree canopy rather than a totally artificial tree. Furthermore, this hybrid tree canopy was created to mimic the roughly symmetrical tree canopy conditions of a field-grown, two-year guava canopy. This design allows for a more realistic assessment of the tree’s growth patterns, canopy physiology, and environmental interactions against spray distribution characteristic’s, providing valuable insights into how guava trees respond to spray droplets dynamics. By simulating field conditions, the study aims to enhance our understanding of tree physiology and optimize cultivation practices with respect to tunnel spraying system. Additionally, the moisture stability within the used guava leaf was not maintained in the study for two reasons: one, the physiology of the hybrid tree might change if replacing the old guava leaves with new ones can skew the performance results, and second, guava leaves were hard enough, and did not lose the moisture significantly and maintained their intact shape.

Detailed description of different adjustment mechanisms

The nozzle angle adjustment mechanism consisted of an upper link, a lower link, a nozzle holder, and an angle adjuster nut and bolt, as provided in Fig. 3a. The upper and lower links were hinged together, but the lower link can move forward and backward in horizontal plane, can’t up and down in vertical plane. However, upper link can move up and down in vertical plane with help of long bolt through a slot to be provided in the lower link. The upper end of long bolt was welded with the upper link, but lower end can move up and down in the slot. To set desired nozzle angle, the hollow cone nozzle was placed in the nozzle holder, welded on the upper link’s surface, which was varied, fixed the nozzle angle using a long bolt and measured as illustrated in Fig. 3b. This separate mechanism was mounted for each nozzle on only one side of each shield.

Fig. 3 [Images not available. See PDF.]

Nozzle angle adjustment mechanism (a) and nozzle angle measurement (b) (1: nozzle, 2: nozzle holder, 3: upper link, 4: long bolt).

In order to supply air volume behind the nozzle, the air deflecting mechanism was fixed to the square opening of the air duct, which was consisted of three MS strips (L×W: 80 × 60 mm), assembled in U shape and used to adjust same orientation with nozzles.

In order to achieve different nozzle spacing as per each experimental run, a nozzle space adjustment mechanism was designed, consisted of an L-shaped angle 750 mm long (35 × 35 × 5 mm) and two flat links to be bolted with only one end of each shield. One face of L-shaped angle was bolted with a tunnel shield, while on the other face, two slots (L×W: 300 × 10 mm) were provided to achieve upward and downward movement of the flat links, as shown in Fig. 4a. One end of these flat links was bolted to the L shaped angle, while another end was bolted to the nozzle angle mechanism. The required nozzle spacing was achieved through upward and downward movement of the flat links, which was further attached with hollow cone nozzles. The simple line diagram for the nozzle spacing adjustment mechanism was shown in the Fig. 4b.

Fig. 4 [Images not available. See PDF.]

Image (a) and simple line diagram (b) of nozzle spacing adjustment mechanism (1: L-shape angle, 2: flat link, 3: lower link of angle adjustment mechanism) (all dimensions, mm).

The tunnel opening adjustment mechanism was fabricated using an L-shaped angle. It consisted of two separate frames, i.e., the shield mounting frame and the blower mounting frame. The linkage of the shield mounting frame was hinged from the top of one tunnel shield, while the blower mounting frame was fixed with a test trolley set up, a centrifugal blower unit, and another tunnel shield, as shown in Fig. 5. Three slots were made in the front end along the length of the shield mounting frame. Two holes were drilled in front of the blower mounting frame and used to fix both frames together for varying tunnel openings as per the experimental run.

Fig. 5 [Images not available. See PDF.]

Front view of tunnel opening adjustment mechanism (all dimensions, mm).

Fin pitch defines the gap between two adjacent fins, i.e., it is maintained by using a fin separator in this study. A fin separator is made of long flat (L × W × T: 1200 × 35 × 4 mm) and fin holder strips (H × W × T: 60 × 60 × 1.5 mm). The fin holder strips were welded over the long flat with the desired fin pitch (20, 40, 60, 80, and 100 mm) by maintaining a 4 mm spacing between these two fin strips, as given in Fig. 6. The four fin separators for each fin pitch were fabricated, fixed each shield and used for each experimental fin pitch.

Fig. 6 [Images not available. See PDF.]

Fin separators for 20 mm fin pitch (all dimensions, mm).

Selection of structural component configurations for efficient recycling system

In this study, four constructional parameters having five level each by considering six center points as suggested by the Design of Expert software were selected based on available research literatures^{31,52, 53–54}. The experimental plan for structural component configurations is given in Table 1.

Table 1. Experiment plan of structural components configuration and responses.

S. no.	Independent parameters	Levels	Dependent parameters
1	Nozzle spacing, mm	5	1. Spray deposition, µg/mm² 2. Recycling rate, %
2	Nozzle spray angle, degree	5
3	Tunnel opening, mm	5
4	Fin pitch, mm	5

Experimental design matrix for structural component configurations

After selection of the range of all constructional parameters, five levels of each parameter were determined, and an experimental design matrix was prepared by considering six center points to reduce random error in the experiment. In the present study, the central composite rotatable design was chosen as given in Table 2. However, an RSM coded values for the various levels of independent parameters are mentioned in Table 3.

Table 2. Experimental design matrix of four input parameters under central composite rotatable design (CCRD).

Runs	Actual values of variables				Responses
Runs	X1	X2	X3	X4	Y1	Y2
1	500	45	1225	60	1.899	29.81
2	400	25	1225	60	1.237	32.36
3	350	55	1100	80	2.601	33.63
4	400	45	1225	60	2.446	32.53
5	350	55	1100	40	2.699	39.98
6	450	35	1350	80	1.578	19.26
7	400	45	975	60	2.944	42.36
8	350	35	1350	40	1.803	25.45
9	450	55	1350	40	1.449	17.27
10	400	65	1225	60	1.702	23.9
11	350	35	1100	80	2.402	36.36
12	450	55	1100	40	2.153	34.81
13	300	45	1225	60	2.621	32.32
14	400	45	1225	60	2.142	36.24
15	400	45	1225	60	2.184	33.95
16	400	45	1225	20	2.392	33.63
17	350	55	1350	80	1.916	18.07
18	450	35	1350	40	1.598	23.63
19	450	55	1100	80	2.180	32.72
20	400	45	1225	60	2.528	31.54
21	450	35	1100	80	2.296	37.19
22	350	35	1100	40	2.534	37.27
23	400	45	1225	100	2.432	20.36
24	450	55	1350	80	1.929	16.81
25	400	45	1225	60	2.453	30.63
26	350	55	1350	40	1.809	25.23
27	400	45	1225	60	2.592	28.18
28	450	35	1100	40	1.772	29.54
29	350	35	1350	80	1.719	23.63
30	400	45	1475	60	1.016	12.27

X1: nozzle spacing (mm); X2: nozzle spray application angle (degree); X3: Tunnel opening (mm); X4: fins pitch (mm); Y1: spray deposits and Y2: recycling rate.

Table 3. Coded value for different levels of input parameters.

Coded values	X1	X2	X3	X4
− 1	350	35	1100	40
0	400	45	1225	60
+ 1	450	55	1350	80
+ α	500	65	1475	100
− α	300	25	975	20

Experimental procedure

The stepwise procedure for each experimental run and response sample collection is given below.

The tunnel sprayer was mounted with a test trolley set up by keeping the desired tunnel spacing between two shields and placing a soil tray in between the two shields, i.e., laid parallel to the direction of travel, as shown in Fig. 7.
Two hybrid trees with the desired number of leaves as per 1.6 LAI were fixed in an erected position in the soil tray by keeping a constant distance of 2 m between them. To ensure free movement of tunnel sprayers, these trees were fixed at the center of both tunnel shields and soil tray.
The desired values of nozzle spacing, nozzle angle, tunnel opening, and fin pitch, respectively, for each experimental run were regulated using suitable adjustments.
The water-sensitive papers (WSP) were fixed at desired positions in three zones of both trees (within 200 mm inward from the beginning of the canopy for both zones, i.e., right and left), as shown in Fig. 8. For each experimental run, a total 36 WSP were used.
The outlets of the recycling basins of both shields were connected to a separate bucket via pipe for recycling samples.
An average air velocity of 15 m/s and an operating pressure of 0.98 MPa were maintained with suitable arrangements.
The tunnel sprayer moved over an hybrid tree canopy with the help of a test trolley set up to maintain a constant forward speed of 0.98 m/s using a flux vector drive arrangement on the basis of the preliminary trial, because system was tested on four different speed (0.69, 0.83, 0.98 and 1.11 m/s), and found 0.98 m/s most suitable speed for achieving maximum spray deposits and minimum recycling rate. However the speed of operation is relatively slow in real world, but study’s objective is not only to achieve higher speed operation but also to achieve higher spray deposits to secure better crop yield by effective handling of insect pests and save environment by recovery of surplus spray liquid. Hence, 0.98 m/s operational speed was chosen for the all experimental runs because it offered higher spray deposits and minimum recycling rate. In this study, the spraying system was immediately stopped when it reached the far end of the trolley set-up to minimize chance of double spray delivery over the tree canopy during the returning time.
The recycling sample was collected from both tunnel shields and used for recycling rate estimation.
All WSP were removed carefully and dried to prevent the overlapping of spray droplets.
These WSP images were digitized and processed in ImageJ software (ImageJ 1.38x, NIH, USA); spray deposits results for all images were recorded.
This whole procedure was repeated for all experimental runs as per the CCRD design matrix as given in Table 2.

Fig. 7 [Images not available. See PDF.]

A prototype tunnel sprayer on test trolley set up (1: Tunnel shield, 2: Artificial tree, 3: Square pole, 4: U channel support, 5: U channel beam (8 m long each), 6: Sprocket, 7: motor, 8: Pulley, 9: roller chain, 10 & 11: rectangular hollow beam, 12: Blower mounting frame, 13: Flat plate, 14: Plastic wheel, 15: Pipe holder, 16: Spray liquid pipe and 17: Direction of travel).

Assessment procedure for response parameters

Assessment of recycling rate (%)

The prototype tunnel sprayer, fitted with four nozzles, moved over simulated hybrid trees from one end to the other end of the test trolley set up (one pass). The recycled sample for each experiment was collected for the pass, measured both recycled liquid and total discharge liquid of four nozzles at same forward speed. Using Eq. (1), recycling rate was estimated for each experimental run^31,32.

where, is recycled liquid, m³/s and is total spray mixture delivered, m³/s and Y2 is recycling rate, %.

Assessment of spray deposits

For the assessment of spray deposits, both simulated hybrid tree canopy was categorized into three zones, such as the left, center, and right zones. Before conducting the experiment, one WSP size of (76 × 26 mm) was cut into three equal parts and fixed on the adaxial and abaxial leaf surfaces of the natural leafs at different positions in the various zones of the both trees using plastic clips (Total WSP: 36 on both trees per experimental run). However, the WSP different locations in single simulated tree canopy were depicted by using yellow and black strips (Fig. 8); a similar pattern was also followed for other hybrid tree canopies.

Fig. 8 [Images not available. See PDF.]

Position of WSP on single simulated tree (top view: a) and (side view, parallel to direction of travel: b).

The tunnel sprayer, fitted with all adjustments, moved over this hybrid tree canopy with the help of a test trolley set up. For each experiment, all WSP were carefully taken from both trees, and dried for sufficient time to prevent droplet dispersion or overlap on these papers. These WSP samples were scanned using a scanner (HP Scanjet G3110), digitized (600 dpi), and then saved. After calibration of the ImageJ software, each WSP image was imported, and processed in the ImageJ software^55,56recorded their spray deposits respective values using following steps as:

To calibrate the software, an RGB image including WSP with plane scale was first imported.
Set the threshold limit by selecting following options “Image > Adjust > Threshold” options.
To set the global scale, draw a straight line on the scale using the line tool. Then, enter the value of the drawn line’s known distance and unit.
Once, the calibration is over, then there is no requirement for further calibration for other WSP images.
The RGB image of WSP was loaded and transformed to a binary image through following “Image > Type > 8 bit” options.
Select the rectangle-shaped tool and select at least 30% of the WSP area from the center of the each imported image to avoid the influence of hands on each side of the image while collecting the from tree canopy, click on USDA automatic paper analysis tool, and again click on drawn box in imported image”.
In this study, the spray liquid density was 997 kg/m³ (the density of water), and the calculation assumed that the spray droplets were spherical in shape.
The spray deposits of the designated region (µg/cm²) are shown.
Spray deposits were then converted to SI units (µg/mm²).
Repeat steps 5 to 8 for every WSP and note the values corresponding to each WSP.

After the assessment of spray deposits for all WSP images, the mean spray deposits on the whole leaf for each zone of both trees was calculated by the average value of the spray deposits of all WSP in that particular zone. The mean spray deposits of all zones were compared to identify a zone that had the minimum mean spray deposits in both trees. The purpose for this comparison is to find out worst zones of both trees, which might be safe haven for insect pests and disease occurrence, finally used mean spray deposits of the worst zones of both trees for optimization to improve efficiency of the tunnel sprayer.

Statistical analysis

The mean spray deposits dataset of different zones of both trees were applied to the Bonferroni post-hoc test (OriginPro, 2021) to compare their means for spatial spray distribution within the canopy. On the basis of minimum spray deposits, the central zones of both trees were selected and their mean deposits dataset was used for further analysis.

The significance of all constructional parameters on responses was determined by analysis of variance (ANOVA) as per the best-fit model. Each response was fitted to a regression prediction model to establish the empirical relationship between the constructional parameters and each response. The general regression model for each response is given in Eq. (2).

where, Y is response variable, β₀ is constant, β_i is linear coefficients, β_ii is coefficients of square order, β_ij is coefficients of interaction terms.

The optimization of four selected constructional parameters was done against means spray deposits in the worst zone and recycling rate. The optimized values of these constructional parameters was achieved by fixing the goal for maximization of mean spray deposits and minimization of recycling rate by keeping all inputs in range, as given in Table 4. The reason for maximization of mean spray deposits and minimization of recycling rate is to ensure greater degree of crop protection from insect pests and disease infestation. However in this study, the surplus applied mixture in terms of recycling rate was harnessed in order to protect environment degradation with effective management of the insect pests. As per goal criteria, numerical optimization was accomplished in the Design of Expert software at constant values of LAI (1.6), forward speed (0.98 m/s), air velocity (15 m/s), and operating pressure (0.98 MPa).

Table 4. Constraints for optimization of structural component configurations with responses.

Parameter	Goal	Lower limit	Upper limit	Center point
Tunnel spacing, mm	Range	975	1475	1225
Nozzle spacing, mm	Range	300	500	400
Nozzle angle, degree	Range	25	65	45
Fins pitch, mm	Range	20	100	60
Y1, 10^− 4 µg/mm²	Maximum	1.016	2.944	2.592
Y2, %	Minimum	12.27	42.36	28.18

Y1: mean spray deposits (µg/mm²) and Y2: recycling rate (%).

Prediction model using hybrid PSO- ANN approach

An artificial Neural Network (ANN) with a multilayer feed-forward back-propagation architecture was developed to model the performance of a recycling tunnel sprayer. Designed in MATLAB 2019b, the network utilized tan-sigmoid activation functions in the hidden layers and linear activation in the output layer. The Levenberg-Marquardt algorithm was employed for efficient training. The ANN architecture consisted of seven input neurons, a hidden layer with 7 and 13 neurons, and six output neurons, forming a 3-7-13-3 structure. A total of 30 data sets were used, with 90% allocated for training and 10% for testing. The model achieved a high coefficient of determination (R² = 0.972), reflecting strong predictive accuracy. Further validation using R², Root Mean Square Error (RMSE), and the Ratio of Performance to Deviation (RPD) confirmed the model’s reliability and precision in Eqs. (3), (4) and (5), respectively. The high R² and low RMSE values indicated a strong correlation between actual and predicted outputs with minimal prediction error.

where, N represent the number of datasets, and denote the actual and predicted output values of the i^th set respectively, and represents the mean of actual output values.

Constructional parameters optimization using hybrid ANN and PSO

In MATLAB 2019b, a hybrid model integrating an Artificial Neural Network (ANN) with the Particle Swarm Optimization (PSO) algorithm was developed to optimize the constructional parameters of a recycling tunnel sprayer. Since the PSO has shown superior performance for modelling complex nonlinear relationships^57,58 for various agricultural machinery operations, and navigating efficiently complex optimization landscapes, as highlighted by various researchers^{57,59, 60–61}. Integrating PSO and other optimization techniques with ANNs enables the identification of optimal operating parameters, thereby enhancing the precision, efficiency, and reliability of agricultural machinery. This integrated approach holds promise for future advancements in smart agricultural technologies. The optimization workflow for this integrated approach is illustrated in the Fig. 9.

Fig. 9 [Images not available. See PDF.]

Flow chart for optimization using hybrid ANN PSO.

The study initiated with training a 3-7-13-3 feed-forward back-propagation artificial neural network (ANN) using experimental data. Once validated for reliability, the ANN was employed to model the constructional parameters of a recycling tunnel sprayer. In the subsequent phase, these parameters were optimized using an enhanced Particle Swarm Optimization (PSO) algorithm. Two PSO variants were explored: a standard PSO with constant inertia weight (ω) and an Improved PSO featuring a linearly decreasing inertia weight and a confined search space to achieve better optimization and faster convergence. The selection of PSO parameters focused particularly on the role of inertia weight. A hybrid ANN-PSO approach was implemented, integrating the Improved PSO method as outlined by researcher⁶¹to improve optimization accuracy. This technique was executed in MATLAB 2010, with the PSO guided by a fitness function defined as the sum of squared errors between the required and ANN-predicted parameters, effectively enhancing the recycling tunnel sprayer’s performance as shown in Eq. (6).

Where F denotes the fitness function, and represents the required and predicted values of the i^th output parameter, respectively, and O is the total number of output parameters to be taken as 3 in this study. In the Improved Particle Swarm Optimization (PSO) algorithm, a swarm consisting of 100 particles is evolved over 1000 iterations. The cognitive acceleration coefficients c1and c2 are both set to 2. The inertia weight factor ω is linearly decreased from 0.9 to 0.4 throughout the iterations. Additionally, the positions of the particles are restricted within predefined boundary limits.

Results

Spatial spray distribution uniformity over simulated hybrid trees canopy

The effect of canopy depth on mean spray deposits of whole leaf for both trees is shown in Fig. 10. Non-significant difference was found between the mean spray deposits of both trees at each depth zone for the whole leaf, except in the central zone of both trees (p < 0.05). However, minor deflection in spray deposits corresponding to the different zones of both trees had been observed due to the orientation of the WSP to be placed in these zones. In addition, a significant difference was found between the mean deposits of the central zone and other zones (left and right) of both trees (p < 0.05), similar results for spray coverage was recorded in previous study⁶². The mean spray deposits at the central zone was recorded at 36.95 and 29.95 per cent less than the mean deposits on the left and right zones of trees 1, respectively, while it registered at 31.58 and 43.73 per cent less than the left and right zones of trees 2, respectively. It could be happened due to the higher canopy density acted as an obstruction to spray droplets⁶³left or right zones are closer to nozzles, drifting of spray droplets during transportation at the center of both simulated trees^64,65as results the central zone always remains away from the nozzles of both shields. On other hand, the left and right zones of both trees directly come into contact with mounted nozzles on both shields. However, in this research, the lowest mean spray deposits is observed in the central zone of both trees and influenced the spatial spray deposition uniformity, could lead to safe haven for the growth of the insect pests and diseases dissemination. The spatially uniform spray dispersion strategy has a wide range of implications for pest management policy, with both potential advantages and disadvantages. The benefits include being less expensive and faster than more specialized pest management techniques, as well as treating the entire area, which lessens the possibility of missing pest populations in particular areas. Because the procedure is simple, governments and agencies may find it easier to control and enforce uniform spraying protocols. On the other hand, spatially uniform spray distribution can lead to inefficient pest management because pests may concentrate in specific areas, pesticide resistance, non-target effects, chemical runoff, resource waste, human exposure risk, and public pushback from environmental groups, local communities, or organic farmers worried about the ecological impact or health risks. The spatially uniform spray dispersion strategy has a wide range of positive implications for pest management policy including being less expensive and faster than more specialized pest management techniques, as well as treating the entire area, which lessens the possibility of missing pest populations in the particular areas. However, government and agencies may find it easier to control and enforce uniform spraying protocols on account of the simple procedure for spatial uniform spray distribution.

Therefore, there is a need to improve the spray deposits at the central zone of the both tree canopy to obtain spatial spray distribution uniformity within the tree canopy. Thus, the mean spray deposits of whole leaf was calculated by averaging the spray deposits of the central zones of both trees, and used for regression model prediction, ANOVA, and optimization analysis to improve the spray deposits the central zones.

Fig. 10 [Images not available. See PDF.]

Canopy depth wise spray distribution for both trees (T1: tree 1 and T2: tree 2), means deposition with same letter represents no statistically significant difference at p < 0.05 (Bonferrni Post-hoc test).

Selection of best fit model for mean spray deposits and recycling rate

The standard criteria were considered for selecting the best-fit model i.e. significant model p-value, non-significant lack of fit, difference between adjusted R² and predicted R² should be less than 0.2, and higher R² values.

In the present scenario, the linear and quadratic models for each response meet these three pre-requisite conditions, as given in Table 5. Hence, the fourth condition, i.e., the R² value of both models for each response were compared, and the selected quadratic model. Since the higher R² value indicates smaller differences between the observed value and the fitted value of the response, it also shows the higher accuracy and predictability of the developed response model.

Table 5. Selection of best fit model for mean spray deposits and recycling rate (%).

Source	Model p-value	Lack of fit	Adj. R²	Pred. R²	R²	Response
Linear	< 0.0001**	0.3463^NS	0.8120	0.7757	0.84	Y1
Quadratic	0.0486^*	0.5366^NS	0.8662	0.7012	0.93	Y1
Linear	< 0.0001**	0.1009^NS	0.5704	0.4737	0.63	Y2
Quadratic	0.0001^**	0.5428^NS	0.8456	0.6562	0.92	Y2

NS, non-significant; Adj., adjusted; Pred., predicted; Y₁: recycling rate; Y₂: mean spray deposits; *: 5% level of significance; **: 1% level of significance.

ANOVA and prediction model for mean spray deposits (Y1) and recycling rate (Y2)

After choosing the quadratic model for each response, an ANOVA was performed by considering experimental data of each mean spray deposits and recycling rate response to determine the significance of each independent parameter on these responses, as given in Table 6. The statistical model for both mean spray deposition and recycling rate was found to be significant (p < 0.01). All parameters were found to be statistically significant on mean spray deposition and recycling rate, except for fin pitch (X4) for mean spray deposits. For mean spray deposits, it was found that the tunnel opening parameters and nozzle angle were significant at the quadratic level. Similar findings have been obtained regarding the significance of the tunnel opening and fin pitch parameters on recycling rate. All interactions effect were found to be non-significant at p < 0.05 against both mean spray deposits and recycling rate.

Table 6. ANOVA for mean spray deposition and recycling rate.

Source	df	Y1		Y2
Source	df	F-value	p-value	F-value	p-value
Model	14	12.34	< 0.0001**	14.41	< 0.0001**
X1	1	19.44	0.0005**	6.05	0.0266*
X2	1	4.76	0.0455*	5.11	0.0390*
X3	1	93.08	< 0.0001**	160.85	< 0.0001**
X4	1	0.9628	0.3420^NS	9.58	0.0074**
X1 × 2	1	0.0185	0.8937^NS	0.0397	0.8448^NS
X1 × 3	1	2.41	0.1414^NS	0.0480	0.8296^NS
X1 × 4	1	2.74	0.1185^NS	2.34	0.1469^NS
X2 × 3	1	0.0927	0.7649^NS	1.92	0.1862^NS
X2 × 4	1	0.0961	0.7609^NS	2.24	0.1551^NS
X3 × 4	1	0.0485	0.8287^NS	1.19	0.2923^NS
X1²	1	0.8940	0.3594^NS	0.3456	0.5653^NS
X2²	1	43.23	< 0.0001**	3.89	0.0672^NS
X3²	1	8.64	0.0102*	5.56	0.0324*
X4²	1	0.0187	0.8931^NS	6.30	0.0241*
Residual	15
Lack of fit	10	0.9842	0.5428^NS	0.9966	0.5366^NS
Pure error	5
Cor total	29

Y1: Mean spray deposits; Y2: Recycling rate; X1: Nozzle spacing, X2: Nozzle angle, X3: Tunnel opening and X4: Fins pitch, *: 5% level of significance, **: 1% level of significance, NS, non-significant; df, degree of freedom.

An empirical relationship was established for both responses, such mean spray deposits and recycling rate. Therefore, a set of new coefficient estimates for significant parameters was produced by discarding the non-significant parameters from the selected quadratic model and used to establish an empirical relationship between constructional parameters and mean spray deposits as well as recycling rate as given in Eqs. 7 and 8.

where, Y1: spray deposits (10^− 4 µg/mm²); Y2: recycling rate (%); X1: nozzle spacing (mm); X2: nozzle angle (degree); X3: tunnel opening (mm) and X4: fins pitch.

Effect of structural component configurations on mean spray deposits

In a 3-D surface plot, the slope of the surface indicates the behavior of each independent (nozzle spacing and nozzle angle) parameter in the spray deposits (Fig. 11a). The effect of nozzle spacing (X1) was found to be significant on mean spray deposits decreased with increasing nozzle spacing. Since small nozzle spacing facilitates narrow spray swath and greater overlapping of spray materials, which forced to pass through deep canopy penetration by the air assistance. The response factor varied from 2.42 to 2.11 × 10^− 4 µg/mm² within selected nozzle spacing, and decreased by 14.69 per cent of the lower value. Additionally, nozzle spacing can have a significant influence on pesticide application distribution and efficacy. Proper nozzle spacing ensures uniform coverage of the target area, reducing the likelihood of missed spots or excessive overlap. This, in turn, enhances the efficacy of the delivered pesticide while minimizing waste and potential environmental impact.

Initially, mean spray deposits was increased with increasing nozzle angle (X2) from 25 to 45 degrees, and then decreased from 45 to 65 degrees (Fig. 10a). The maximum spray deposits were found to be 2.96 × 10^− 4 µg/mm² at a 45-degree nozzle angle. In addition, nozzle angle also influences spray coverage and drop properties, specifically regarding the distribution pattern and droplet size. A steeper angle may produce a narrower spray pattern with finer droplets, while a more obtuse angle can create a wider coverage with larger droplets, impacting both efficiency and effectiveness in various applications. It is also a critical for achieving optimal performance in various applications, such as agricultural spraying and industrial cleaning. By adjusting the nozzle angle, operators can enhance efficiency and ensure that the intended surfaces receive adequate treatment.

Fig. 11 [Images not available. See PDF.]

Effect of nozzle spacing and nozzle angle (a) and fin pitch and tunnel opening (b) against mean spray deposits (Y1: 10^− 4 µg/mm²).

The input parameters, namely tunnel opening and fin pitch, were plotted against mean spray deposits in a 3-D surface plot (Fig. 11b). The slope of the 3-D surface graph presented a typical relationship between these input parameters and spray deposits.

The tunnel opening (X3) was found to have a statistically significant effect on mean spray deposits. The mean spray deposits decreases with increasing tunnel opening. Since small tunnel opening offer close exposure between vertically aligned set of nozzles within shields and tree canopy, lead to greater chance for deep canopy penetration of fine stream of spray droplets, result into higher spray deposits and vice versa. The highest spray deposits were registered at 1100 mm tunnel opening. With tunnel opening variation from 1100 to 1350 mm, mean spray deposits was decreased of 1.57 times of the greater value.

In this study, the effect of the fin pitch (X4) was found to be non-significant against mean spray deposits (Fig. 11b). It was varied 3.02 to 2.87 × 10^− 4 µg/mm², decreased only 1.05times of the higher deposits value, because there is not direct relationship between fin pitch and the spray deposits.

Effect of structural component configurations on recycling rate (%)

In the 3-D surface graph, the results of recycling rate were plotted against the different input values of nozzle spacing and nozzle angle (Fig. 12a). This study registered an inverse relationship between recycling rate and nozzle spacing (X1). Since minimum nozzle spacing provided lower overlapping of the spray material and narrow spray swath, which had higher scope to be captured by the recycling system, it suggests that optimizing nozzle spacing could enhance the efficiency of material recovery. Consequently, further research may be necessary to explore the potential benefits of adjusting nozzle configurations in various recycling processes. Recycling rate varied from 35.82 to 29.65% corresponding to selected range of nozzle spacing and decreased by 6.17%.

Initially, the recycling rate was increased with increasing nozzle angle (X2) from 25 to 45 degrees and then decreased after 45 degrees (Fig. 12a), because as droplets are moved closer to attain projectile path from 25 to 45 degrees, they proceed with greater velocity and move to towards projectile motion, as results recycling rate are improving. At 45 degrees, this angle allows for a more concentrated delivery of materials, ensuring that the droplets reach their intended collection points with minimal loss. After 45 degree, they propagate away from the projectile path, the droplets may lose their intended trajectory, causing them to scatter and miss the target areas designated for collection. Consequently, optimizing nozzle angles is crucial to maintaining an effective recycling process while maximizing material recovery. In this study, recycling rate ranged from 35.82 to 41.82 per cent and 40.82 to 37.33 per cent with varying nozzle angles from 25 to 45 degrees and 45 to 65 degrees, respectively.

Fig. 12 [Images not available. See PDF.]

Effect of nozzle spacing and nozzle angle (a) and forward speed and tunnel opening (b) against recycling rate (Y2: %).

Figure 12b presents the effect of tunnel opening (X3) on recycling rate. The recycling rate (%) results showed an inverse relationship with tunnel opening. Since a small tunnel opening facilitates a small distance between two shields, i.e., moved over the tree canopy, a set of nozzles to be mounted vertically in each shield delivered the stream of spray droplets over the tree canopy, a significant portion of the delivered pesticide retained by the tree canopy, whereas surplus pesticide passed through the tree canopy, struck the opposite shield, and collected as a recycling rate. In case of a large tunnel opening, a lesser quantity of applied pesticide reached the opposite shield due to the large existing gap between both shields, which reduced the spray material collection in the recycling form. It varied from 39.43 to 18.06% within the selected range of tunnel openings and decreased by 21.37%.

Generally, there is an inverse relationship between recycling rate and fin pitch. Since a lower fin pitch makes it possible to place more fins in each shield, increasing the area available for capturing spray droplets and decreasing their reflection off each shield. As a result, a pattern of inverse proportionality between recycling rate and fin pitch was noted. The recycling rate was improved from 20 to 40 mm fin pitch (X4) and then decreased for further fin pitch range (Fig. 12b). It increased from 39.43 to 41.20 per cent and decreased from 41.20 to 34.53 per cent with a fin pitch range of 20 to 60 mm and 60 to 100 mm, respectively. However, with a fin pitch of 20 mm, lower recycling rates were recorded than with 40 mm. It might have happened because the lowest fin pitch allowed the installation of the highest number of fins and acted as a reflection surface rather than a droplets capturing surface, resulting in decreased recycling efficiency. Consequently, optimizing fin pitch is crucial for improving the overall performance of tunnel sprayer.

Optimization of independent parameters against responses

The best optimum solution for the desired goal was selected on the basis of a high desirability function and desired response attributes, as given in Table 7. Since the desirability function indicates the percentage of achieving the intended attributes of the goal set. At 0.85 desirability functions, the optimized values of independent parameters offered 2.75 × 10^− 4 µg/mm² of mean spray deposits and saved 38.80 per cent of applied spray mixture, and confirmed greater degree of crop protection with sustainable inputs.

Table 7. Optimum solutions for structural components configuration.

Parameter	Optimum value
Nozzle spacing, mm	350.00
Nozzle angle, degrees	42.50
Tunnel spacing, mm	1100.00
Fins pitch, mm	41.00
Y₁, µg/mm²	2.75 × 10^− 4
Y₂, %	38.80
Desireability	0.85

Y₁: mean spray deposits (µg/mm²) and Y₂: recycling rate (%).

Optimization of constructional parameters of a recycling tunnel sprayer using hybrid PSO-ANN algorithm

The optimization of constructional parameters for a recycling tunnel sprayer was accomplished by coupling the best-performing prediction model with Particle Swarm Optimization (PSO). The PSO was configured with an inertia weight of 0.9, acceleration coefficients of 2, a swarm size of 100, and 1000 iterations. This setup enabled effective exploration of performance metrics. As shown in the convergence curve (Fig. 13), the algorithm approached optimal solutions by iterations 987 and 989 for spray deposits and recycling rate, respectively, with the mean fitness values closely matching the global best.

Table 8 presents the optimal parameters derived from the ANN-PSO model: nozzle spacing (X1) of 354 mm, nozzle angle (X2) of 42 degree, tunnel opening (X3) of 1142 mm, and fin pitch (X4) of 41 mm. The integration of prediction models with optimization algorithms demonstrates an advanced strategy for improving recycling tunnel sprayer efficiency. The convergence analysis confirms the model’s reliability, highlighting PSO’s ability to fine-tune parameters for maximizing spray deposition and recycling efficiency.

Table 8. Optimized performance parameters by ANN-PSO.

X1, mm	X2, degree	X3, mm	X4, mm	Spray deposits, 10^− 4 µg/mm²
X1, mm	X2, degree	X3, mm	X4, mm	Predicted	Observed	Deviation (%)
354	41.7	1122	40.7	95.76	97.23	1.97

X1, mm	X2, degree	X3, mm	X4, mm	Recycling rate, %
X1, mm	X2, degree	X3, mm	X4, mm	Predicted	Observed	Deviation (%)
352	42.0	1078	41.5	96.28	97.43	1.15

Nozzle spacing (X1, mm); nozzle angle (X2, degree); tunnel opening (X3, mm); fin pitch (X4, mm).

Fig. 13 [Images not available. See PDF.]

PSO convergence characteristics on Spray deposits (10^− 4 µg/mm²) and Recycling rate (%).

A thorough analysis revealed that the integrated PSO with ANN strategy outperformed the RSM technique in terms of prediction accuracy for both response parameters with minor variation in optimized constructional parameters. This discrepancy suggests that while the RSM method has its merits, the combination of Particle Swarm Optimization (PSO) with Artificial Neural Networks (ANN) may offer a more robust solution for enhancing prediction accuracy in complex modeling scenarios. Future studies could explore the underlying factors contributing to this difference and seek to improve the RSM methodology.

Discussion

Testing techniques employed in the research to evaluate the tunnel sprayer

It is a tunnel sprayer designed with an integrated tunnel structure, rather than a conventional sprayer that was tested inside a tunnel. The tunnel sprayer incorporates a partially enclosed framework that surrounds the target crop, allowing for controlled spraying within a confined space. This design helps in directing the spray more efficiently toward the plant canopy while simultaneously reducing spray drift and enhancing deposition efficiency. The tunnel structure also facilitates the collection and recycling of excess spray, thereby improving overall application efficiency and minimizing chemical wastage.

The research employed a systematic experimental approach to evaluate the performance of the tunnel sprayer, focusing on the influence of key spray control parameters, including nozzle angle, nozzle spacing, tunnel opening, and fin pitch, on spray deposits and recycling rate. The testing involved controlled trials conducted in a structured experimental setup designed to simulate real-world spraying conditions in an orchard environment. To quantify spray deposits, water-sensitive papers were strategically placed at various canopy positions to capture spray coverage and droplet distribution. These papers were later analyzed using image processing techniques to determine coverage percentage and droplet density.

To evaluate the recycling efficiency, the sprayer was equipped with a collection system that captured excess spray that did not adhere to the target surfaces. The recovered spray volume was measured and compared against the total spray volume applied to determine the recycling rate. The influence of nozzle angle was examined by adjusting the orientation of the nozzles to study its effect on spray penetration and coverage uniformity. Nozzle spacing variations were tested to evaluate how different distances between nozzles affected the spray overlap and distribution within the canopy. Tunnel opening settings were modified to determine their impact on airflow dynamics and spray containment within the spraying zone, while fin pitch adjustments were analysed to understand their role in directing spray movement, enhancing deposits efficiency and recycling rate. The combination of these testing techniques allowed for a comprehensive evaluation of the tunnel sprayer’s performance, providing insights into optimal parameter configurations to maximize spray deposits while minimizing waste through an efficient recycling mechanism.

The nozzle spacing defines the spacing between two consecutive nozzles, it can be horizontal or vertical distance, but vertical spacing was only taken into consideration for this experimental research and horizontal spacing was kept zero, because all nozzles were arranged in vertical straight line. However, the nozzle spacing influences the total swath and the overlapping area to be shared by two adjacent nozzles, and finally influence spray quality distribution into the canopy⁶⁶. Since the small nozzle spacing facilitates narrow spray swaths and higher overlapping of the spray mixture, and vice versa. Therefore, at small nozzle spacing, air assistance disturbed the physiology of the canopy, and fine droplets penetrate deeply into the canopy^67,68,, led to higher spray deposits within tree canopy. In addition, the spacing between small nozzles plays a critical role in determining spray uniformity because nozzle spacing influences the spray dynamics, spray swath, overlapping of spray material and flow rate delivered. If the nozzles are too close together, their spray patterns overlap excessively, leading to over-application in certain areas, which can cause runoff or plant damage. Conversely, if the nozzles are too far apart, gaps may form, leading to under-application and reduced effectiveness. Hence, optimizing the nozzle spacing involves balancing various factors (spray angle, nozzle height, flow rate, operating pressure) to ensure that every section of the target area receives the correct amount of spray without over- or under-application. Several researches have demonstrated that nozzle spacing significantly influences spray uniformity in agricultural applications. For instance, researcher⁶⁹ developed a computer program to determine optimal nozzle spacing based on factors like operating pressure, nozzle height, and spray pattern overlap. The study found that for a Triple Action Nozzle with a hollow cone pattern, operating at 4.0 kg/cm² and a height of 40 cm, the optimal nozzle spacing was approximately 57 cm, resulting in minimal coefficient of variation and improved spray uniformity. Similarly, research conducted on the effects of nozzle geometry, spacing, and height on pesticide spray characteristics⁷⁰. This research finding revealed that nozzle spacing between 50 and 75 cm on a boom combined with a spraying height of 50 cm, produced excellent surface coverage, emphasizing the importance of appropriate nozzle arrangement to prevent spray pockets and ensure uniform distribution. .

The optimized value of the nozzle spacing was recorded of 350 mm, which far less than above cited research work. It might be happened due to difference in physiology of the tree canopy and speed.

The recycling rate decreases with increasing nozzle spacing because large nozzle spacing offers large spray swath with limited tunnel shield capacity. It leads to increase ground losses with drift loss and reduces recycling rate. .

Nozzle orientation with respect to tree canopy influences spray quality parameters i.e. spray penetration⁷¹spray distribution, uniformity coverage⁶² and recovery of applied pesticide, specifically in tunnel sprayer. In the present case, the effect of the upward nozzle angle with horizontal plane was analyzed while the transverse nozzle angle was kept constant as per the previous study⁵². The effect of nozzle angle was observed significant against both spray deposits and recycling rate. In the present scenario, both spray deposits and recycling rate was increased with increasing nozzle angle up to 45 degrees, and then decreased for further nozzle angle. Since, the stream of spray droplets moves towards projectile motion from 25 to 45 degrees and non-projectile motion from 45 to 65 degrees. The tunnel opening concept is confined specifically to the tunnel sprayer, which indicates the horizontal gap between two tunnel shields; these shields moved over the tree canopy. The small tunnel opening offers a minimum gap between both shields, and vice versa. Hence spray opening of all nozzles with air assistance to be mounted in each shield remained close to canopy, offered greater degree of spray droplets penetration, resulted into higher spray deposits into the tree canopy and more recycling rate, similar recycling rate trend were reported by the researcher³¹. Additionally, spray deposits and recycling rate can be improved through air assistance. Since Air assistance alters the tree canopy’s physiology, including the leaves’ orientation, movement, and area density per unit volume. It also exposes the leaf surface to a stream of spray droplets and forces them to enter the tree canopy’s depth. Additionally, it helps the spray create a finer stream of droplets that have a greater range and disperse across the canopy, increasing the recycling rate, spray deposits, and uniformity of spray deposition^{43, 44–45}. This enhanced distribution ultimately improves the effectiveness of pest and disease management strategies, ensuring that treatments reach all necessary areas of the foliage. Moreover, by optimizing the coverage, air assistance contributes to healthier tree growth and increased yield potential.

Recovery of delivered pesticide depends on the shield capture area, thus several fins were laid in the tunnel sprayer to improve recycling rate by increasing capturing area, which can be increased by incorporating more fins maintaining uniform fin pitch. The fin pitch represents the gap between two fins. Thus lower fin pitch permits installation of more fins into each shield, and creates more scope for capturing spray droplets and reduces the reflection of spray droplets from each shield. Hence inversely proportional trend for recycling rate was observed with respect to fin pitch. In addition, the effect of fin pitch against the spray deposits was found non-significant because there is no direct correlation between installed fins and spray droplets. The role of shield fins begin after spray deposits to be occurred within the canopy.

Potentail contribution of tunnel spraying system for environment conservation and future scope with advance technologies

The distinct recycling system of the tunnel sprayer consisted of tunnel shields, which made a tunnel-like structure that moved over the tree canopy within the row. The vertical nozzle’s arrangement is done in each shield in such a way that the spray mixture is thrown from the set of these nozzles of one shield over the tree canopy, and significant spray material is deposited. Whereas, surplus spray material is passed through the canopy, even empty space between trees, and captured by the opposite shield. In this study, the maximum recycling rate was recorded at about 38.80% under optimized conditions and can be reused in the field. This study’s results indicated that the tunnel sprayer had prevented the mixing of harmful chemicals with each environment component and conserved it without compromising the degree of crop protection.

In the future, the tunnel spraying system can be tested with use of a solar energy by mounting solar plate on the back of each shield. For instance, several researchers had adopted various techniques to improve the use of solar energy⁷². It would help to transform the tunnel spraying system into more eco-friendly technology. The study’s findings can influence the current tunnel sprayer’s sturdiness, compact design and aid in future enhancements to increase its efficacy and efficiency. Additionally, study outcome contribute to scaling up the farmer’s income by effective handling of insect pests with minimum inputs, and extending the use of the optimized system to other agricultural crops with certain modifications. Additionally, the findings could pave the way for innovative agricultural practices that enhance crop yield while reducing chemical dependency, ultimately fostering a more sustainable approach to farming. By implementing these advancements, farmers may also benefit from reduced labor costs and increased productivity, leading to a more resilient agricultural sector.

From a comprehensive review of literature on tunnel sprayers, it reveals that the tunnel spraying system was not tested comprehensively with the association of advanced technologies. In the recent past, various research works related to these emerging technologies have been conducted for effective and advanced spraying in agriculture, like machine learning^{73, 74–75}artificial intelligence^76,77sensors^78,79electrostatic spray mechanisms^{80, 81–82}and mathematical prediction models⁸³etc. Hence, there is a tremendous scope for integration of advanced technologies with the tunnel spraying system for achieving precise and effective application of delivered pesticide by analyzing data in real-time. These technologies can improve the next level of accuracy and efficiency for identifying weeds and insect pests, dynamically adjusting based on real-time conditions, and predicting maintenance for monitoring equipment performance and predicting failures before they happen.

In this study, the effect of four constructional parameters was evaluated against spray deposits and recycling rate, and their significant effect is discussed in detail within the result section. Various researchers had evaluated the effect of several factors to improve the performance of the tunnel sprayer, such as type of nozzles, air outlet orientation, forward speed, air velocity, and tunnel opening, but the majority of research work was carried out within vineyard canopies. However, there is still a need to investigate the effects of several constructional and operational parameters against tunnel sprayer performance parameters, including fin pitch, air duct diameter, air outlet type and size, and operating pressure. Additionally, the effect of climatic characteristics (ambient temperature and relative humidity) on the increased concentration of the active ingredient of the applied spray mixture in the actual field condition has to be investigated and the need to create a mathematical model based on these climatic factors to neutralize the effect of the increased concentration of active ingredients.

Conclusions

The post-hoc test showed the uniform spatial distribution of spray deposits over the tree canopy, except central zones of the both trees, which depicted the accuracy and potential ability of the tunnel sprayer for comprehensive canopy coverage with minimum environment’s quality degradation. In this research, the effect of all constructional parameters was found to be statistically significant against recycling rate as well as mean spray deposits, except fin pitch in the case of mean spray deposits. It revealed that the adjustment of these constructional parameters in the tunnel spraying system is necessary for achieving better spray deposits with high spray efficacy and additional benefits of considerable recycling rate. However the spray deposits was improved by 14.73% and 22.87% by increasing nozzle angle from 34 to 45 degree and decreasing nozzle spacing from 450 to 350 mm, respectively. Recycling rate was improved by 21.37% and decreased by 6.17% within selected range of tunnel opening and nozzle spacing, respectively.

With a set of optimized values of constructional parameters, mean spray deposits was found of 2.75 × 10⁴ µg/mm² with high desirability function (0.85). It assured greater degree of crop protection with surplus saving of 38.80 per cent of applied spray mixture, which authorizes to the tunnel sprayer as an eco-friendly and precision based effective and efficient spraying technique for sustainable use of inputs. Moreover, the integration of Particle Swarm Optimization with Artificial Neural Networks provided a more robust solution for enhancing prediction accuracy in both responses. Since this synergy leverages the strengths of both techniques, allowing the swarm intelligence of PSO to effectively tune the parameters of ANNs.

Acknowledgements

Authors are thankful to IARI for financial support in scholarship form and to ICAR-CIAE for providing the laboratory facilities. There is no specific grant issued for the research.

Author contributions

J.S., M.D., K.N.A. and B.J. (Conceptualization and designing, Methodology ; J.S., K.N.A., B.J., A.K.R. (Supervision, investigation, Execution of experiments); J.S., B.J., A.K.R., J.M. and V.P. (Data curation); J.S., M.D., B.J. (Software, Formal analysis and data interpretation); J.S., M.D., K.N.A. (Writing- review & editing).

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

Data is provided within the manuscript.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

This study confirmed that there is no harm physically or mentally to animal as well as human during experimentations.

Consent to participate

All authors have agreed and given approval for the submission. We confirm that this manuscript has not been published elsewhere and is not also under consideration by any another journal.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Chambers, JE et al. Human and ecological risk assessment of a crop protection chemical: A case study with the Azole fungicide Epoxiconazole. Crit. Rev. Toxicol.; 2014; 44, pp. 176-210.1:CAS:528:DC%2BC2cXitVylsrg%3D [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24274332]

2. Guthman, J; Brown, S. I will never eat another strawberry again: The biopolitics of consumer-citizenship in the fight against methyl iodide in California. Agric. Hum. Values; 2016; 33, pp. 575-585.

3. Pimentel, D; Levitan, L. Pesticides: Amounts applied and amounts reaching pests. Bioscience; 1986; 36, pp. 86-91.1:CAS:528:DyaL28XhtVChs7s%3D

4. Pimentel, D. Amounts of pesticides reaching target pests: Environmental impacts and ethics. J. Agric. Environ. Ethics; 1995; 8, pp. 17-29. [DOI: https://dx.doi.org/10.1007/BF02286399]

5. Fox, RD; Derksen, RC; Zhu, H; Brazee, RD; Svensson, SA. A history of air-blast sprayer development and future prospects. Trans. ASABE; 2008; 51, pp. 405-410. [DOI: https://dx.doi.org/10.1016/j.cropro.2014.05.006]

6. Dekeyser, D. et al. Performance assessment of orchard sprayers, Part 1: Machine Characterization. in Proceedings 11th Workshop on Sustainable Plant Protection Techniques in Fruit Growing, 34–35 (2011).

7. TriloffP. Results of measuring the air distribution of sprayers for 3D-crops and parameters for evaluating and comparing fan types. 13th Workshop Spray. Application Fruit Grow. (Suprofruit) 21–23. https://doi.org/10.5073/jka.2015.448.000 (2015).

8. Triloff, P. Results and conclusions from five years measuring and adjusting air distribution of brand new sprayers for 3D-Crops. in 6th European Worksh op on Standard Procedure for the Inspection of Sprayers in Europe (SPISE) 13–15 (2016).

9. Holterman, HJ; van de Zande, JC; Huijsmans, JF; Wenneker, M. An empirical model based on phenological growth stage for predicting pesticide spray drift in pome fruit orchards. Biosyst. Eng.; 2017; 154, pp. 46-61. [DOI: https://dx.doi.org/10.1016/j.biosystemseng.2016.08.016]

10. Blanco, et al. Real-time monitoring of spray drift from three different orchard sprayers. Chemosphere; 2019; 222, pp. 46-55.2019Chmsp.222..46B1:CAS:528:DC%2BC1MXhvFOitLc%3D [DOI: https://dx.doi.org/10.1016/j.chemosphere.2019.01.092] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30690400][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6472945]

11. Rathnayake et al. Downwind spray drift assessment for airblast sprayer applications in a modern apple orchard system. Trans. ASABE64, 601–613. https://doi.org/10.13031/trans.14324 (2021).

12. Kirk, IW; Hoffmann, WC; Carlton, JB. Aerial electrostatic spray system performance. Trans. ASAE; 2001; 44, 1089.

13. Pascuzzi, S; Cerruto, E. Spray deposition in tendone vineyards when using a pneumatic electrostatic sprayer. Crop Prot.; 2015; 68, pp. 1-11.

14. Yamane, S; Miyazaki, M. Charging performance of an electrostatic pesticide spraying system for low-concentration and high-volume applications. J. Jpn.Soc. Agric. Mach.; 2008; 70, pp. 115-122.

15. Ru Yu, RY; JiaQiang, Z; Zhou, ZJ; HongPing, ZH; Shu ChaoRan, SC. Design and experiment of double-nozzle of aerial electrostatic sprayer. Trans. CSAM; 2007; 38, pp. 58-61.

16. Martini, AT et al. Pesticide drift from aircraft applications with conical nozzles and electrostatic system. Ciência Rural; 2016; 46, pp. 1678-1682.1:CAS:528:DC%2BB3MXktlKlsL0%3D

17. Patel, MK et al. An advance air-induced air-assisted electrostatic nozzle with enhanced performance. Comput. Electron. Agric.; 2017; 135, pp. 280-288.

18. Maghsoudi, H; Minaei, S; Ghobadian, B; Masoudi, H. Ultrasonic sensing of pistachio canopy for low-volume precision spraying. Comput. Electron. Agric.; 2015; 112, pp. 149-160. [DOI: https://dx.doi.org/10.1016/j.compag.2014.12.015]

19. Bietresato, M; Carabin, G; Vidoni, R; Gasparetto, A; Mazzetto, F. Evaluation of a LiDAR-based 3D-stereoscopic vision system for crop-monitoring applications. Comput. Electron. Agric.; 2016; 124, pp. 1-13. [DOI: https://dx.doi.org/10.1016/j.compag.2016.03.017]

20. Shen, Y; Zhu, H; Liu, H; Chen, Y; Ozkan, E. Development of a laser-guided, embedded-computer-controlled, air-assisted precision sprayer. Trans. ASABE; 2017; 60, pp. 1827-1838. [DOI: https://dx.doi.org/10.13031/trans.12455]

21. Li, et al. A canopy density model for planar orchard target detection based on ultrasonic sensors. Sensors; 2016; 17, 31.2016Senso.17..31L1:CAS:528:DC%2BC2sXjtlegsbo%3D [DOI: https://dx.doi.org/10.3390/s17010031] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28029132][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5298604]

22. Cheraiet, A et al. An algorithm to automate the filtering and classifying of 2D lidar data for site-specific estimations of canopy height and width in vineyards. Biosyst. Eng.; 2020; 200, pp. 450-465. [DOI: https://dx.doi.org/10.1016/j.biosystemseng.2020.10.016]

23. Baio, FHR et al. Variable rate spraying application on cotton using an electronic flow controller. Prec. Agric.; 2018; 19, pp. 912-928. [DOI: https://dx.doi.org/10.1007/s11119-018-9564-7]

24. Anifantis, AS; Camposeo, S; Vivaldi, GA; Santoro, F; Pascuzzi, S. Comparison of UAV photogrammetry and 3D modeling techniques with other currently used methods for estimation of the tree row volume of a super-high-density olive orchard. Agriculture; 2019; 9, 233. [DOI: https://dx.doi.org/10.3390/agriculture9110233]

25. del-Moral-Martínez, I; Rosell-Polo, JR; Uribeetxebarria, A; Arnó, J. Spatially variable pesticide application in vineyards: Part I, developing a geostatistical approach. Biosyst. Eng.; 2020; 195, pp. 17-26. [DOI: https://dx.doi.org/10.1016/j.biosystemseng.2020.04.014]

26. Campos, J; García-Ruíz, F; Gil, E. Assessment of vineyard canopy characteristics from vigour maps obtained using UAV and satellite imagery. Sensors; 2021; 21, 2363.2021Senso.21.2363C [DOI: https://dx.doi.org/10.3390/s21072363] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33805351][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8036331]

27. Ade, G; Molari, G; Rondelli, V. Vineyard evaluation of a recycling rate tunnel sprayer. Trans. ASAE; 2005; 48, pp. 2105-2112. [DOI: https://dx.doi.org/10.13031/2013.20087]

28. Baldoin, C., De Zanche, C. & Bondesan, D. Field testing of a prototype recycling sprayer in a vineyard: Spray distribution and loss. Agric. Eng. Int.: CIGR J.10 (2008).

29. Thériault, R; Salyani, M; Panneton, B. Development of a recycling rate sprayer for efficient orchard pesticide application. Appl. Eng. Agric.; 2001; 17, pp. 143-150. [DOI: https://dx.doi.org/10.13031/2013.5456]

30. Pergher, G. & Petris, R. A novel, air-assisted tunnel sprayer for vineyards. in International Conference Innovation Technology to Empower Safety, Health and Welfare in Agriculture and Agro-food Systems September, 15–17 (2008).

31. Pergher, G; Petris, R. A novel, air-assisted tunnel sprayer for vineyards: Optimization of operational parameters and first assessment in the field. J. Agricultural Eng.; 2009; 40, pp. 31-38. [DOI: https://dx.doi.org/10.4081/jae.2009.88]

32. Cerruto, E., Manetto, G., Balloni, S. & Caruso, L. A recycling rate tunnel to reduce environmental drift in spraying goblet vineyards. in International Conference Safety, Health and Welfare in Agriculture and Agro-Food Systems, September, 3–6 (2012).

33. Bozzer, C., Tamagnone, M. & Balsari, P. Performance Evaluation of Recycling rate Sprayer in Vineyard. in International Workshop on Vineyard Mechanization and Grape and Wine Quality 978, 199–204 (2013).

34. Siegfried, W. & Holliger, E. Application Technology in Fruit-Growing and Viticulture. Report, Swiss Federal Research Station (Wädenswill, Switzerland, 1996).

35. Viret, O; Siegfried, W; Holliger, E; Raisigl, U. Comparison of spray deposition and efficacy against powdery mildew of aerial and ground-based spraying equipment in viticulture. Crop Prot.; 2003; 22, pp. 1023-1032. [DOI: https://dx.doi.org/10.1016/S0261-2194(03)00119-4]

36. Panneton, B; Lacasse, B; Piché, M. Effect of air-jet configuration on spray coverage in vineyards. Biosyst. Eng.; 2005; 90, pp. 173-184. [DOI: https://dx.doi.org/10.1016/j.biosystemseng.2004.11.001]

37. Ilker, H. C., Mehmet, R. D., rkan, G. & Erdal, K. G. A. Effect of air assistance on deposition distribution on spraying by tunnel-type electrostatic sprayer. Afri. J. Agric. Res.4, 1392–1397. https://doi.org/10.5897/AJAR.9000196 (2009).

38. Furness, G. O. & Pinczewski, W. V. A comparison of the spray distribution obtained from sprayers with converging and diverging airjets with low volume air assisted spraying on citrus and grapevines. J. Agric. Eng. Res.32, 23–10. https://doi.org/10.1016/0021-8634(85)90095-2 (1985).

39. Pergher, G; Gubiani, R; Tonetto, G. Foliar deposition and pesticide losses from three air-assisted sprayers in a hedgerow vineyard. Crop Prot.; 1997; 16, pp. 25-33. [DOI: https://dx.doi.org/10.1016/S0261-2194(96)00054-3]

40. Pezzi, F; Rondelli, V. The performance of an air-assisted sprayer operating in vines. J. Agric. Eng. Res.; 2000; 76, pp. 331-340. [DOI: https://dx.doi.org/10.1006/jaer.2000.0540]

41. Walklate, P. J., Weiner, K. L. & Parkin, C. S. Computational fluid dynamics: a tool for evaluating air assistance sprayer design concepts. in Second international symposium on pesticide application techniques 101–107 (1993).

42. Molari, G., Benini, L. & Ade, G. Design of a recycling rate tunnel sprayer using CFD simulations. Trans. ASAE 48, 463–468. https://doi.org/10.13031/2013.18309 (2005).

43. Cross, J. V., Walklate, P. J., Murray, R. A. & Richardson, G. M. Spray deposits and losses in different sized apple trees from an axial fan orchard sprayer: 3. Effects of air volumetric flow rate. Crop Prot.22, 381–394 (2003). https://doi.org/10.1016/S0261-2194(02)00192-8

44. Pergher, G. Improving vineyard sprayer calibration - air flow rate and forward speed. Annual Rev. Agricultural Eng.; 2005; 4, pp. 197-204.

45. Ade, G; Molari, G; Rondelli, V. Recycling rate tunnel sprayer for pesticide dose adjustment to the crop environment. Trans. ASABE; 2007; 50, pp. 409-413. [DOI: https://dx.doi.org/10.13031/2013.22632]

46. Jamar, L et al. Comparative performance of recycling rate tunnel and conventional sprayers using standard and drift-mitigating nozzles in dwarf apple orchards. Crop Prot.; 2010; 29, pp. 561-566. [DOI: https://dx.doi.org/10.1016/j.cropro.2009.12.018]

47. Ambrogetti, A. O. et al. Evaluation of a drift recovery panel for phytosanitary treatments in vineyards. Revista de la Facultad de Ciencias Agrarias, Universidad Nacional de Cuyo48, 83–94 (2016). https://revista.fca.uncu.edu.ar/images&.

48. Anifantis, A. S., Pascuzzi, S. & Santoro, F. Vineyard Treatments Performed With a Recycling rate Tunnel Sprayers Prototype: Preliminary Assessment. in IX International Scientific Symposium Farm Machinery and Processes Management in Sustainable Agriculture, Lublin, Poland (2017).

49. Pergher, G; Zucchiatti, N. Influence of canopy development in the vineyard on spray deposition from a tunnel sprayer. J. Agric. Eng.; 2018; 49, pp. 164-173. [DOI: https://dx.doi.org/10.4081/jae.2018.801]

50. Singh, J. et al. Estimation of leaf area and leaf area density for design optimization of a recycling tunnel sprayer. J. Sci. Ind. Res.81, 173–179 (2022). http://nopr.niscpr.res.in/handle/123456789/59081

51. Singh, J et al. Assessment of morphological attributes of guava tree for design of a novel tunnel spraying system. Environ. Conserv. J.; 2024; 25, pp. 1-9.

52. Pergher, G et al. Assessment of spray deposition and recycling rate in the vineyard from a new type of air-assisted tunnel sprayer. Crop Prot.; 2013; 45, pp. 6-14. [DOI: https://dx.doi.org/10.1016/j.cropro.2012.11.021]

53. Holownicki, R., Doruchowski, G. & Świechowski, W. Uniformity of spray deposit within apple tree canopy as affected by direction of the air-jet in tunnel sprayers. J. Fruit Ornament. Plant Res.5, 129–136 (1997).

54. Holownicki, R., Doruchowski, G. & Świechowski, W. Minimising pesticide waste and emissions to the environment by using tunnel sprayers. J. Fruit Ornament. Plant Res.5, 137–144 (1997).

55. Zhu, H; Salyani, M; D Fox, R. A portable scanning system for evaluation of spray deposition distribution. Comput. Electron. Agric.; 2011; 76, pp. 38-43. [DOI: https://dx.doi.org/10.1016/j.compag.2011.01.003]

56. Mangado, J et al. Measuring the accuracy of a pesticide treatment by an image analyzer. Proc. Technol.; 2013; 8, pp. 498-502. [DOI: https://dx.doi.org/10.1016/j.protcy.2013.11.066]

57. Anantachar, M., Kumar, P. G. & Guruswamy, T. Neural network prediction of performance parameters of an inclined plate seed metering device and its reverse mapping for the determination of optimum design and operational parameters. Comput. Electron. Agric., 72, 87–98. https://doi.org/10.1016/j.compag.2010.03.001 (2010)

58. Pareek, C. M., Tewari, V. K., Machavaram, R. & Nare, B. Optimizing the seed-cell filling performance of an inclined plate seed metering device using integrated ANN-PSO approach. Artif. Intell. Agricu.5, 1–12. https://doi.org/10.1016/j.aiia.2020.11.002 (2021).

59. Pareek, C. M., Tewari, V. K. & Machavaram, R. Multi-objective optimization of seeding performance of a pneumatic precision seed metering device using integrated ANN-MOPSO approach. Eng. Appl. Artif. Intell.117, 105559. https://doi.org/10.1016/j.engappai.2022.105559 (2023).

60. Sawant et al. Optimization of operational parameters of Site-specific pesticide spray module for young pomegranate orchards using RSM and RBFNN-PSO techniques. J. Biosyst. Eng.48, 225–241 (2023).

61. Sankar, P. A., Machavaram, R. & Shankar, K. System identification of a composite plate using hybrid response surface methodology and particle swarm optimization in time domain. Measurement55, 499–511. https://doi.org/10.1016/j.measurement.2014.05.025 (2014).

62. Dai, S et al. Effects of sprayer speed, spray distance, and nozzle arrangement angle on low-flow air-assisted spray deposition. Front. Plant Sci.; 2023; 14, 1184244. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37223814][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10200916]

63. Duga, AT et al. Spray deposition profiles in pome fruit trees: Effects of sprayer design, training system and tree canopy characteristics. Crop Prot.; 2015; 67, pp. 200-213. [DOI: https://dx.doi.org/10.1016/j.cropro.2014.10.016]

64. Grella, M; Miranda-Fuentes, A; Marucco, P; Balsari, P. Field assessment of a newly-designed pneumatic spout to contain spray drift in vineyards: Evaluation of canopy distribution and off-target losses. Pest Manage. Sci.; 2020; 76, pp. 4173-4191.1:CAS:528:DC%2BB3cXhsFShurbO [DOI: https://dx.doi.org/10.13031/trans.57.10646]

65. Rathnayake, AP et al. Downwind spray drift assessment for airblast sprayer applications in a modern apple orchard system. Trans. ASABE; 2021; 64, pp. 601-613. [DOI: https://dx.doi.org/10.13031/trans.14324]

66. Griesang, F., Spadoni, A. B. D. & Ferreira, P. H. U. & Da Costa ferreira, M. Effect of working pressure and spacing of nozzles on the quality of spraying distribution. Crop Prot.151, 105818 (2022).

67. Triloff, P. A case study-canopy adapted dosing and spray application in 3D crops. Dose Expr Workshop Assoc. Appl. Biol 6–7 (2018).

68. Bernat, S et al. Use of ultrasound anemometers to study the influence of air currents generated by a sprayer with an electronic control airflow system on foliar coverage. Effect droplet size. Comput. Electron. Agric.; 2022; 202, 107381. [DOI: https://dx.doi.org/10.1016/j.compag.2022.107381]

69. Samuel, A., Mingxiong, O., Weidong, J., Eric, A. A. & Wanting, Y. Assessing nozzle geometry, spacing and height effect on pesticide spray characteristics and swath from ground and aerial sprayers. Eur. Sci. J.16, 30. https://doi.org/10.19044/esj.2020.v16n30p103 (2020).

70. Sayinci, B; Demir, B; Acik, N. Comparison of spray nozzles in terms of spray coverage and drop distributionuniformity at low volume. Turk. J. Agric. For.; 2020; 44, pp. 262-270.1:CAS:528:DC%2BB3cXisFCjsbzE

71. de Lima, J. I. D. S., Degrande, P. E., de Souza, C. M. A. & Viegas neto, A. L. Droplet distribution as a function of cotton interrow spacing and angles of sprayer displacement. Engenharia Agricola38, 927–933. https://doi.org/10.1590/1809-4430-eng.agric.v38n6p927-933/2018 (2018).

72. Arulprakasajothi, M; Saranya, A; Srimanickam, B; Devarajan, Y; Raja, ND. Study on optimizing the energy gradient and temperature regulation of flat plate solar collectors with advanced hybrid nanofluids. J. Therm. Anal. Calorim.; 2024; 149, pp. 13309-13328.1:CAS:528:DC%2BB2cXhvV2msLnE [DOI: https://dx.doi.org/10.1007/s10973-024-13555-0]

73. Acharya, P; Burgers, T; Nguyen, K. A deep-learning framework for spray pattern segmentation and Estimation in agricultural spraying systems. Sci. Rep.; 2023; 13, 7545.2023NatSR.13.7545A1:CAS:528:DC%2BB3sXhtVSjt7rL [DOI: https://dx.doi.org/10.1038/s41598-023-34320-7] [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37160930][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10170126]

74. Jose, B; Hampp, F. Machine learning based spray process quantification. Int. J. Multiph. Flow; 2024; 172, 104702.1:CAS:528:DC%2BB2cXjs1Glug%3D%3D [DOI: https://dx.doi.org/10.1016/j.ijmultiphaseflow.2023.104702]

75. Eberle, M et al. Application of machine learning for the prediction of particle velocity distribution and deposition efficiency for cold spraying titanium powder. Comput. Mater. Sci.; 2024; 244, 113224.1:CAS:528:DC%2BB2cXhsF2ru7bM [DOI: https://dx.doi.org/10.1016/j.commatsci.2024.113224]

76. Sharma, S; Verma, K; Hardaha, P. Implementation of artificial intelligence in agriculture. J. Comput. Cogn. Eng.; 2023; 2, pp. 155-162. [DOI: https://dx.doi.org/10.47852/bonviewJCCE2202174]

77. Talaviya, T; Shah, D; Patel, N; Yagnik, H; Shah, M. Implementation of artificial intelligence in agriculture for optimisation of irrigation and application of pesticides and herbicides. Artif. Intell. Agric.; 2020; 4, pp. 58-73. [DOI: https://dx.doi.org/10.1016/j.aiia.2020.04.002]

78. Abbas, I et al. Different sensor based intelligent spraying systems in agriculture. Sens. Actuators A: Phys.; 2020; 316, 112265.1:CAS:528:DC%2BB3cXit1yjt7vF [DOI: https://dx.doi.org/10.1016/j.sna.2020.112265]

79. Wang, Y et al. A review of environmental sensing technologies for targeted spraying in orchards. Horticulturae; 2025; 11, 551. [DOI: https://dx.doi.org/10.3390/horticulturae11050551]

80. Lin, Z et al. Research and experiment of electrostatic spraying system for agricultural plant protection unmanned vehicle. Front. Ecol. Evol.; 2023; 11, 1138180.

81. Salcedo, R et al. Evaluation of an electrostatic spray charge system implemented in three conventional orchard sprayers used on a commercial apple trees plantation. Crop Prot.; 2023; 167, 106212. [DOI: https://dx.doi.org/10.1016/j.cropro.2023.106212]

82. Guo, Z et al. Study on deposition characteristics of the electrostatic sprayer for pesticide application in greenhouse tomato crops. Agriculture; 2024; 14, 1981. [DOI: https://dx.doi.org/10.3390/agriculture14111981]

83. Shani, B. B. Development of a mathematical model for forecasting spray deposition on tree canopies. Dutse J. Pure Appl. Sci. (DUJOPAS) 7. https://doi.org/10.1016/j.icheatmasstransfer.2010.02.003 (2021).

Word count: 12212

Show less

corrected publication 2025. This work is published under http://creativecommons.org/licenses/by-nc-nd/4.0/ (the "License"). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.

Abstract

Translate

A tunnel sprayer has distinct recycling systems, which save a significant portion of the applied pesticide. Its performance is directly or indirectly influenced by various constructional parameters of the recycling system (nozzle spacing, nozzle angle, tunnel opening, and fin pitch) because they influence the movement of spray droplets, characteristics of the droplets, and recycling rate of the system. Therefore, the effect of these parameters was investigated for improvement in the tunnel sprayer’s performance. In this study, four adjustment mechanisms were developed for selected parameter’s regulation and tested using a hybrid tree (artificial and natural leaf). The non-significant results of spatial spray distribution revealed its accuracy and potential ability for maintaining spray uniformity across the tree canopy, except in the central zone of both trees. By considering the poor spray deposits at the central zone of both trees, the statistical analysis was carried out for spray deposits maximization and recycling rate minimization. The spray deposits varied from 1.016 × 10^− 4 to 2.944 × 10^− 4 µg/mm² and offered better plant protection. The sprayer saved up to 42.36% of the applied pesticide, which ensures economical farm operations and saves the environment. The optimized values for input parameters (i.e., X1: 350 mm, X2: 42.5 degrees, X3: 1100 mm, and X4: 41 mm) would improve the efficiency of the tunnel sprayer and conserve costly input resources.

Details

Title

Enhancing tunnel sprayer performance through optimization of constructional parameters

Author

Singh, Jagjeet¹

; Din, Maharani²; Agrawal, Kamal Nayan³; Jyoti, Bikram¹; Roul, Ajay Kumar¹; Mahapatra, Jyotirmay¹

; Pagare, Vikas¹

¹ ICAR-Central Institute of Agricultural Engineering, Nabi Bagh, 462038, Bhopal, Madhya Pradesh, India (ROR: https://ror.org/026j5b854) (GRID: grid.464528.9) (ISNI: 0000 0004 1755 9492)
² AICRP on ESA, ICAR-CIAE, Bhopal, India; Draught Animal Power, Farm Machinery and Related Aspects, RRA Network/WASSAN, Hyderabad, India
³ AICRP on FIM, ICAR-CIAE, Bhopal, India

Pages

20349

Section

Article

Publication year

2025

Publication date

2025

Publisher

Nature Publishing Group

e-ISSN

20452322

Source type

Scholarly Journal

Language of publication

English

DOI

https://doi.org/10.1038/s41598-025-07194-0

ProQuest document ID

3228193317

Enhancing tunnel sprayer performance through optimization of constructional parameters

Jump to:

Full text

Abstract

Details

Suggested sources