Numerical simulation of downwash airflow distribution inside tree canopies of an apple orchard from a multirotor unmanned aerial vehicle (UAV) sprayer

Highlights

• A CFD model for predicting downwash airflow inside tree canopies was developed using computational fluid dynamics.

• The wireless simulation parameter measurement system (WSPM-System) was used to provide real input data of the rotor speed for numerical simulation.

• Most of the simulated values agreed well with the measured values, and the R2 values obtained were 0.84 and 0.89.

• The airflow velocity inside tree canopies was highly and synthetically influenced by five factors of load, operation height, canopy diameter, tree height, and canopy densities.

Abstract

The downwash airflow from an unmanned aerial vehicle (UAV) sprayer has an important effect on the deposition and penetration of droplets inside canopies. Although UAV sprayers have been widely adopted in commercial spray scenarios in China, the study of downwash airflow inside tree canopies is still lacking. In this study, a CFD model for predicting downwash airflow inside tree canopies was developed using computational fluid dynamics (CFD). The tree canopies were defined as the porous medium in the computational domain. The wireless simulation parameter measurement system (WSPM-System) was used to provide real input data of the rotor speed for numerical simulation. The developed CFD model was validated in two steps through the measurement experiments of the vertical downward velocity (V_VD) of the downwash airflow with trees (EXP-With-Tree) and without trees (EXP-Without-Tree). For the validation of cases with trees and without trees, the R² values obtained were 0.84 and 0.89, respectively, which meant that most of the simulated values agreed well with the measured values. The validated CFD model was used to predict downwash airflow distribution inside tree canopies of typical cases with various application parameters (load, operation height), tree dimensions (canopy diameter, tree height), and canopy densities. The application results showed that the airflow velocity inside tree canopies was highly and synthetically influenced by five factors but could not be determined by any and only one of them. The developed CFD model will be beneficial to make a better understanding of the effect of application parameters and tree structures on the distribution of downwash airflow inside tree canopies and provide references for the UAV sprayers application in orchards.

Introduction

Chemical control of pests and diseases is an integral activity in modern conventional agriculture and contributes to increased yield and quality (Ahmad et al., 2020, Soheilifard et al., 2020). However, only a fraction of the total amount of plant protection products (PPP) is deposited on the intended target, especially during spray applications in high growing crops (such as apple orchard) with conventional sprayers (Garcerá et al., 2017). According to public information (FAO, 2018), in 2018, the total amount of pesticide applications in China was 177.37 × 10⁴ tons, accounting for 43.03% of the world’s total pesticide applications. Moreover, the total pesticide emissions of China increased from 165.47 tons in 2004 to 179.77 tons in 2017 (Jiao et al., 2020). The excessive use of pesticides and off-target losses have raised concerns about human health risks (Kim et al., 2017) from pesticides and contamination of water and soil (Pullan et al., 2016, Sybertz et al., 2020).

Over the past few decades, although air-assisted sprayers have been developed and used for spray pesticides on nurseries and orchards (Walklate et al., 1996, Qi and Fu, 1998, Gu et al., 2012, Grella et al., 2020), the hand-held gun hydraulic sprayer is still used by some growers in China due to their comparatively low cost of purchase and operation. In addition, air-assisted sprayers and other ground-vehicle sprayers are not suitable for orchards with special terrains, such as hills and mountains. The traditional spray method is inefficient and easily causes workers to be exposed to more pesticides, which is extremely hazardous to their health. In recent years, a new spray application technology based on the unmanned aerial vehicle (UAV) sprayer has developed rapidly, particularly in East Asian countries (such as China, South Korea, and Japan) (He et al., 2017) and achieved good application results in field crops of rice paddies (Xue et al., 2014), corn (Zheng et al., 2017), wheat (Wang et al., 2019) and cotton (Chen et al., 2021). However, it is worth noting that the application technique of UAV sprayers in field crops has become increasingly mature, more and more manufacturers and service organizations have paid their attention to the applications of economic crops with higher additional value, especially orchards (Wang et al., 2021a). The reduced exposure of workers to pesticides makes the UAV sprayer applications more attractive (Jorge et al., 2020).

Currently, some attempts have been made to evaluate the spray application effect of UAV sprayers in orchards (Changling et al., 2020, Hou et al., 2019, Jorge et al., 2020, Li et al., 2021, Meng et al., 2020, Tang et al., 2018, Wang et al., 2021a, Zhang et al., 2016, Li et al., 2018). On one hand, these studies were mainly exploring the selection of the spraying parameters, such as flight altitude, flight speed, etc. On the other hand, the quality of the application and economic costs with those of conventional treatment (air-assisted sprayer) was compared and analyzed. Jorge et al., (2020) revealed that the capacity in hectares per hour of the UAV sprayer was greater than conventional equipment in olive and citrus orchards and the application cost was not much different. In addition, the data (coverage, residue, and penetration) collected from one of the most difficult pest control systems (almond orchards) in US agriculture supports reported successes experienced in China (Li et al., 2021). These previous studies provide some fundamental guidance for the spray application of UAV sprayers in orchards. However, researchers seldom focus on distribution characteristics of downwash airflow inside and around the target tree canopy. Similar to the horizontal and/or vertical airflow from air-assisted sprayers, the downwash airflow of the UAV sprayer plays an important role in the transportation of droplets from their source (nozzles) to the target (Hao et al., 2019a, Hao et al., 2019b). Strong airflow can increase on-target deposition by reducing droplet flight time and reducing the effect of meteorological conditions on the spray (Hong et al., 2018a). Inevitably, the downwash airflow can change the shape of plants (Li et al., 2019, Shi et al., 2021, Tian et al., 2021) and affect the deposition and penetration of droplets inside the crop canopy. Different sources of airflow can lead to different spatiotemporal distribution characteristics for airflow and thus insights into the spatiotemporal distribution of downwash airflow is an important prerequisite in the analysis of droplets deposition dynamics. To improve the spraying efficacy of UAV sprayers and reduce off-target losses, it is essential to understand the distribution characteristics of downwash airflow inside and around the target tree canopy.

In the agricultural sector, two main methods have been used for studying downwash airflow: field measurements using anemometers and numerical simulations with computers. Although the particle image velocimetry (PIV) technology and the lidar-based wind scanner has been used to measure the blade tip vortex (Bauknecht et al., 2017) and the two-dimensional downwash airflow (Sjöholm et al., 2014) from a helicopter, respectively, it is still quite difficult to measure the downwash airflow in complex agricultural scenarios. Because there are complex and three-dimensional interactions between the downwash airflow from the UAV sprayer and the crop canopy. Li et al., (2019) found that the airflow velocity at the bottom of the rice canopy has the greatest attenuation, indicating that the dense leaves at the lower part hinder the downwash airflow. Shi et al., (2021) showed that the degree of rice plant deformation was closely related to the maximum velocity of the downwash airflow. Moreover, Tian et al., (2021) found that the type and flight parameters of the UAV sprayer determine the distribution characteristics of the interference area (the area where the downwash airflow causes the crop canopy to deform), which is the final position affecting droplets deposition on the crop. The data from field measurements (Hu et al., 2014, Chen et al., 2017, Tang et al., 2019) not only provided a reference for the reasonable selection of UAV sprayer types and application parameters but also revealed the interaction between the field crop canopy (rice and wheat) and the downwash airflow. However, due to uncontrollable environmental conditions (ambient wind speed magnitude and direction, etc.) and limited by test costs, test equipment, and test methods, it is not yet possible to obtain the complete spatiotemporal distribution of downwash airflow of UAV sprayers through field measurements (Zhang et al., 2020). The numerical simulation technology can effectively solve the above problems and complement field measurements. To this effect, the approach has become a mainstream research tool for the downwash airflow from UAV sprayers.

The effect of the fruit tree canopy on the spatiotemporal distribution of downwash airflow is significant (Hao et al., 2019a, Hao et al., 2019b). However, it is worth noting that previous numerical simulation studies of downwash airflow (Yang et al., 2017, Wen et al., 2019, Guo et al., 2020, Hao et al., 2020, Tang et al., 2020, Tang et al., 2021) rarely included the fruit tree canopy. For early computational fluid dynamics (CFD) studies on air-assisted sprayers, the fruit tree canopy was modeled in two methods. The first method was that geometric of full-scale trees, including all the branches, were modeled as a three-dimensional (3D) object. Porous sub-domains were also added around the branches to simulate thin branches and leaves (Endalew et al., 2009, Endalew et al., 2010, Duga et al., 2017). The second method was that tree canopies were modeled as a porous medium with cuboidal or spherical geometries (Mercer, 2009, Hong et al., 2018a, Hong et al., 2018b). The former studies took advantage of more computational time and cost to offer a realistic representation of tree canopies, while the latter might strike a balance between numerical simulation and economy. According to the second method, Hong et al., (2018c) replaced the fruit tree canopy with a porous medium for CFD simulation and developed the simulation of air-assisted sprayers (SAAS) application software using the simulation data, which can predict the sprayer's droplet drift in orchards based on user input conditions. The above studies have fully proved that it is feasible to model the fruit tree canopy with the porous medium.

The objective of this study was to develop a CFD model to simulate the downwash airflow inside tree canopies when a six-rotor UAV sprayer in the hovering spray scenario. The full-scale three-dimensional (3D) model of the UAV sprayer was used in the developed CFD model, and the most important significance was that the tree canopy was designed and it was defined as the porous medium in the computational domain. A specially designed wireless simulation parameter measurement system (WSPM-System) was used to provide real input data of the rotor speed for numerical simulation. The developed CFD model was validated in two steps using the field experimental data. Eventually, the validated CFD model was applied to predict downwash airflow distribution inside tree canopies of typical cases with various application parameters of UAV sprayer, tree dimensions, and canopy densities. This CFD model will be beneficial to make a better understanding of the effect of application parameters and tree structures on the distribution of downwash airflow inside tree canopies and provide the opportunity of predicting deposition and drift of spray droplets when integrated with multiphase models.