Simulation and Optimization of a Self-Cleaning Device for the Header of a Rice Seed Harvester Using Fluent–EDEM Coupling

Abstract

Rice seed production is a critical step in breeding high-quality varieties. To ensure seed purity, it is essential that no residual grains or broken ears remain in the harvester header after harvesting each variety, thus preventing cross-contamination. This study addresses the issue of seed retention in existing rice harvesters, which lack efficient self-cleaning or other cleaning mechanisms and cannot be cleaned rapidly. A self-cleaning device for the harvester header was designed to enable one-click cleaning after harvesting a single variety. A novel cleaning nozzle was developed as the key component of the device, with its structure optimized through single-factor and orthogonal combination experiments. The number of nozzles was determined based on their spray width and the header width. A header-cleaning airflow simulation model based on Fluent–EDEM coupling was constructed to investigate the effects of nozzle inlet pressure, airflow incident angle, and nozzle outlet height on the self-cleaning rate. Optimal cleaning parameters were identified to maximize the self-cleaning rate, and the simulation results were validated. The study revealed that the nozzle’s expansion section length, throat diameter, and contraction section length significantly affect the spray width. When the expansion section length was 10 mm, the throat diameter was 8 mm, and the contraction section length was 8 mm, the nozzle achieved the largest jet angle, measuring 50.3 cm. Further analysis indicated that inlet air pressure had the greatest influence on the self-cleaning rate, followed by airflow incident angle and nozzle outlet height. The optimal parameter combination was identified as an inlet air pressure of 0.6 Mpa, an airflow incident angle of 118.25°, and a nozzle outlet height of 2.64 mm, achieving a maximum self-cleaning rate of 99.63%. A one-click cleaning system was designed using an STM32 microcontroller and hardware circuits. Field experiments under optimal parameters demonstrated a self-cleaning rate of 97.68% with a cleaning duration of 10 s per cycle. The findings provide theoretical guidance for the design and optimization of self-cleaning headers for rice seed production.

1. Introduction

Rice is a major staple crop, with its yield largely dependent on seed quality. The seed industry plays a fundamental role in ensuring national food security, making the improvement of seed quality and the acceleration of new variety development key strategies for increasing agricultural productivity [1,2,3]. The production of rice seeds is an essential step in the dissemination of high-quality varieties. Large-scale and mechanized seed production bases serve as stabilizing forces for ensuring seed supply capacity and safety. Harvesting for seed production requires that no residual grains or broken ears remain in the harvester after completing one variety, to prevent cross-contamination with the next [4]. Currently, there are no specialized harvesters for seed production available in the market. To enhance efficiency, general field harvesters are typically employed [5]. The header, which accumulates the largest quantity of residual grains during harvesting, is often cleaned manually by users after completing one variety. This manual cleaning process is time-consuming and labor-intensive and has limited effectiveness, leading to increased seed losses. Therefore, developing a concealed automatic cleaning header is of great significance for improving the efficiency of seed harvesting.

At present, there are few studies on rice harvesters for breeding operations; therefore, the research available in this field is insufficient. There is a strong correlation between plot breeding equipment and seed production equipment in cleaning. In recent years, several scholars have conducted research on plot harvester cleaning and self-cleaning mechanisms. Zhao Chunhua [6] developed a handheld plot combine harvester, applying principles of techno-economics and functional analysis theory. This harvester featured a comb-type harvesting device that utilized the strong negative-pressure airflow generated by the threshing drum to transport grains, effectively addressing the issue of grain retention during the transportation phase in traditional methods. Dai Fei [7] aimed to further improve the performance of cone-shaped threshing devices in plot breeding. By employing the CFD–DEM coupling method, the motion of materials within the axial cone–cylinder threshing transmission device was numerically simulated, achieving axial transport of grains and short stalks while ensuring no grain retention within the device during harvesting. Shang Shuqi’s team [8] adopted an airflow cleaning approach by installing nozzles on both sides of the header. These nozzles used pneumatic airflow to blow residual grains from the sides of the header to the central conveyor belt, ensuring no grain retention in the header during harvesting. Yao Wenyan [9] analyzed areas prone to grain retention in the header and introduced bottom trapdoors at these locations. Additionally, a reversing gear structure was implemented in the scraper conveyor to facilitate reverse cleaning, thus addressing the retention issue. Xia Shengbo [10] investigated the effects of transverse auger speed and airflow velocity at the seed-cleaning inlet on cleaning performance. By conducting numerical simulations of gas–solid two-phase flow within the grain unloading auger and solving the model using CFD–DEM coupling, the study obtained the velocity distribution and motion trajectories of the airflow and particle phases during the seed-cleaning process. The grain storage part of the rice harvester is not only the header but also the high-level unloading barrel. Bochuan Ding [11] studied the cleaning mechanism of single-channel cleaning devices to minimize grain losses during rice harvesting. Using the Discrete Element Method (DEM) and Computational Fluid Dynamics (CFD), the grain motion patterns within the single-channel cleaning device were elucidated. In the realm of agricultural machinery intelligentization, Zhe Qu [12] addressed the issue of wheat harvesting losses caused by improper parameter settings. A cleaning loss monitoring device was designed using STM32 microcontrollers and piezoelectric sensors, effectively reducing harvest losses. Jiarui Wang [13] tackled the high missed-seeding rates in mechanized potato planting by designing a replanting control system based on an STM32 microcontroller, an electric push rod, and through-beam laser sensors. This system efficiently delivered replanted potatoes into furrows, significantly reducing the missed-seeding rate and improving the intelligent level of potato planters.

This study addresses the issue of seed retention in the key header components of existing rice seed production harvesters, which lack efficient self-cleaning or other cleaning functions and are incapable of rapid cleaning. An automatic header cleaning device was designed with the following primary objectives:

(1) Improving the header structure to enable residual seed recovery and reduce seed losses. An aggregate bin was installed at the bottom of the header near the cutter, equipped with an electric push rod to control the opening and closing of the aggregate bin lid. Additionally, retractable nozzles were designed at the rear of the header.

(2) Optimizing the nozzle structure to ensure no residual seeds in the header. Single-factor and orthogonal experiments were conducted using Fluent 19.0.0 fluid simulation software to optimize the structure of individual nozzles. MATLABR2023b software was used to analyze and fit the velocity contour obtained from the simulations, calculating the maximum spray width of a single nozzle. The installation positions of the nozzles were determined based on the header length to ensure that the airflow could cover the entire header. Using the Discrete Element Method (DEM) coupled with Computational Fluid Dynamics (CFD), EDEM 2023.1 and Fluent software were employed to analyze the interaction between grain particles and the airflow field. Based on the results of single-factor experiments and using the self-cleaning rate as the evaluation criterion, a three-factor, three-level central composite experiment was designed to determine the optimal nozzle parameters.

(3) Designing a control system to achieve one-click cleaning. A system hardware framework was constructed, and control program codes were developed. Field experiments were conducted to validate the system, and the results met the required performance indicators. This study provides a reference for the intelligent cleaning of rice seed production harvesters.

2. Materials and Methods

2.1. Concealed Cleaning Header and Working Principle

2.1.1. Header Structure and Key Cleaning Components

Cleaning is a post-harvest operation; therefore, the cleaning components must not interfere with normal harvesting activities. The self-cleaning header designed in this study allows the key components of the cleaning device to remain concealed during regular harvesting operations, only being deployed during cleaning when the machine is stopped. This design protects the key components from damage while ensuring that harvesting operations are unaffected. The device primarily consists of an aggregate bin, a collection lid, nozzles, and nozzle telescopic tubes. Its structure is shown in Figure 1, and the key components are shown in Figure 2, wherein Figure 2a is the aggregate bin and Figure 2b is the telescopic nozzle.

The aggregate bin is positioned beneath the header and is equipped with an electric push rod. The bottom plate of the header also functions as the lid of the aggregate bin. During cleaning, the electric push rod, controlled by a command, pushes the lid open. The concealed telescopic nozzle tube consists of an inner tube and an outer tube, with the outer tube fixed to the rear baffle of the header and the nozzle mounted on the inner tube. The bottom of the aggregate bin features a movable sliding plate, which can be removed after harvesting a single variety to collect residual materials from the header and reduce seed losses. By adjusting the angle at which the outer tube aligns with the rear baffle of the header, the airflow incident angle can be modified.

2.1.2. Cleaning and Self-Cleaning Device Operation Process

During operation, the combine harvester must stop and lower the header to its lowest position. The driver presses the “one-click cleaning” control button in the cabin, and under the control of the STM32 microcontroller, Shanghai, China, the telescopic push rod hidden in the aggregate bin opens the lid of the aggregate bin. Subsequently, a DC motor drives the nozzles outward via a rack-and-pinion mechanism. Once the nozzles are extended, the solenoid valve opens, and high-speed airflow ejected from the nozzles blows any remaining debris on the header into the aggregate bin. Upon completion of the cleaning program, the nozzles retract, and the lid automatically closes, thus finishing the cleaning of the header. The harvester then proceeds to harvest the next variety.

The airflow during this process is supplied by a 12 V air compressor, which is connected in parallel to the harvester’s 12 V onboard battery. An air pressure regulator adjusts the incident air pressure to different settings.

2.2. Optimization of Cleaning Nozzle Structure and Analysis of Airflow Field

2.2.1. Cleaning Nozzle Structural Design

Figure 3a shows a simplified diagram of the airflow cleaning operation. During cleaning, residual seeds, influenced by gravity, accumulate at the curved bottom plate of the header. Under the influence of high-speed airflow, the seeds are blown into the aggregate bin. The fan-shaped nozzle is a critical component in achieving effective airflow cleaning. The performance of the nozzle must ensure that under certain incident pressures, the nozzle’s spray width is maximized to guarantee optimal self-cleaning effects. The structural design of the nozzle is illustrated in Figure 3b.

The nozzle structure reveals that its performance is influenced by several parameters: the expansion section distance d₁, contraction section distance d₂, throat diameter d₃, and nozzle outlet height e. When optimizing the spray width, existing measurement tools do not provide a clear visualization of how different dimensional parameters affect the spray width. Fluent simulation software, known for its convenience and practicality, is used in this study. By leveraging Fluent for simulation verification, the optimal nozzle structure is designed.

2.2.2. Nozzle Optimization Simulation Experiment

• (1). Single-factor Simulation Experiment Method

To investigate the effects of various parameters such as the expansion section distance, contraction section distance, throat contraction diameter, inlet air pressure, and nozzle outlet height on the spray width, single-factor experiments were conducted for each parameter [14]. The nozzle structure is designed to be flexibly telescopic, with an outlet width set at 30 mm. Prior to the single-factor experiments, the setup included: factor A expansion section distance d₁ at 15 mm, factor B contraction section distance d₂ at 8 mm, factor C throat diameter d₃ at 10 mm, inlet air pressure at 0.2 Mpa, and outlet height e at 2 mm. The arrangement of the single-factor experiments is shown in Table 1.

The three-dimensional model of the nozzle created in SolidWorks was imported into the Fluent module of ANSYS 19.0 The fluid domain extraction was completed in SpaceClaim, followed by mesh generation in the mesh tool. After completing the mesh division, a mesh quality check was conducted to ensure that the mesh quality met the requirements. In the solver, the model with standard wall functions was selected, the fluid was set to air, and the pressure inlet was set at 0.2 Mpa. The solution method chosen was Coupled, and the calculation was performed with 2000 iterations.

Given that Ansys software does not directly provide a specialized tool for measuring angles, this study utilized MATLAB data analysis software to preprocess the velocity contour maps obtained from the simulation [15]. Specifically, median filtering techniques were first applied to reduce image noise, followed by binarization to simplify the image information for easier boundary recognition. Then, an algorithm was used to precisely extract the spray face boundaries from the processed image. Finally, curve fitting was applied to the boundary points, and a program was written to automatically calculate the spray angles [16], thereby achieving precise quantitative analysis of the spray characteristics. The process of calculating the nozzle angle is illustrated in Figure 4.

• (2). Optimization of Nozzle Parameters Using Orthogonal Experiments

Based on the results of the single-factor experiments, when other factors are constant, inlet air pressure and outlet height primarily affect the outlet airflow velocity but have minimal impact on the spray width. However, when the inlet air pressure and outlet height are constant, the expansion section distance, throat contraction distance, and throat diameter significantly influence the spray width. To determine the optimal working parameters for the nozzle, orthogonal experiments were conducted using the spray width as the evaluation criterion [17]. The levels of the experimental factors are shown in Table 2. The factors were defined as follows: factor A was the expansion section distance d₁, factor B was the contraction section distance d₂, and factor C was the throat diameter d₃. Three levels were selected for each factor, keeping other nozzle structural dimensions constant and ignoring the interaction effects among factors. Based on the L₉ (3³) orthogonal experiment table, nine groups of nozzles with different geometric parameters were simulated, and the results of the orthogonal experiments are presented in Table 2.

2.3. CFD-DEM Coupling Simulation for Header Residual Cleaning

In the Fluent–EDEM coupling module, the particle phase and gas phase can be coupled for calculation. The Eulerian–Lagrangian method, based on a single-phase flow framework, considers only the momentum exchange between the gas phase and the particle phase, requiring the solid volume fraction to be less than 10%. In this study, the solid volume fraction in the gas–solid two-phase flow is less than 5%, and the particles have small volume and mass. Therefore, the significant influence of particles on the flow field can be ignored, allowing the use of one-way coupling to improve simulation efficiency [18,19]. Flow field information is computed at specific time intervals using Fluent software. By writing a user-defined function (UDF) and following the description format for particle body forces provided in the EDEM 2021 documentation, this flow field information is exported as text files. These files are then imported into EDEM using its API interface to load the forces acting on the particles in the flow field [20,21] for coupling calculations.

2.3.1. Gas–Solid Two-Phase Flow Mathematical Model

• (1). Gas Phase Governing Equations

In Fluent, the gas–phase flow typically follows the Navier–Stokes equations, which describe the conservation of momentum, mass, and energy in a fluid. For incompressible fluids, and when the energy equation is neglected [22] (isothermal flow), the primary equations include: Mass Conservation Equation (Continuity Equation):

(1)$\mathrm{f}\mathrm{r}\mathrm{a}\mathrm{c}\partial \mathsf{\rho }\partial \mathrm{t}+\nabla \cdot \left(\mathsf{\rho }\mathrm{u}\right)=0$

where ρ is the fluid density, u is the velocity vector of the fluid, and t represents time.

The momentum conservation equation (Navier–Stokes equation) is as follows:

(2)$\mathrm{r}\mathrm{h}\mathrm{o}\left({\displaystyle \frac{\partial \mathrm{u}}{\partial \mathrm{t}}}+\left(\mathrm{u}\cdot \nabla \right)\mathrm{u}\right)=-\nabla \mathrm{p}+\mathsf{\rho }\mathrm{g}+\nabla \cdot \left[\mathsf{\mu }\right(\nabla \mathrm{u}+(\nabla \mathrm{u}{)}^{\mathrm{T}})]-{\mathrm{F}}_{\mathrm{p}}$

where p represents the pressure, g denotes the gravitational acceleration vector, μ indicates the dynamic viscosity of the fluid, Fp refers to the force exerted by the particles on the fluid (in one-way coupling, this force is typically ignored or treated as a source term).

• (2). Solid Phase Governing Equation

In EDEM, the motion of solid particles follows Newton’s second law. The particle motion equation can be expressed as:

(3)$\mathrm{m}\mathrm{a}\mathrm{t}\mathrm{h}\mathrm{b}\mathrm{f}{\mathrm{I}}_{\mathrm{i}}{\displaystyle \frac{\mathrm{d}{\mathsf{\omega }}_{\mathrm{i}}}{\mathrm{d}\mathrm{t}}}=\sum _{\mathrm{j}\ne \mathrm{i}}{\mathrm{T}}_{\mathrm{c}\mathrm{o}\mathrm{l},\mathrm{i}\mathrm{j}}+{\mathrm{T}}_{\mathrm{w}\mathrm{a}\mathrm{l}\mathrm{l},\mathrm{i}}$

2.3.2. CFD Simulation Model

During cleaning, all five nozzles perform identical cleaning actions. To simplify the coupling model and improve calculation accuracy [23,24], this study selects the two central nozzles of the original header as simulation objects. The simplified cleaning model is shown in Figure 5a, with the dimensions of the cleaning device being 9600 mm × 680 mm × 590 mm. The three-dimensional model is saved and imported into SpaceClaim19.0.0 software for fluid domain extraction. Structured mesh generation of the fluid domain is performed using the Workbench Meshing module, as illustrated in Figure 5b. The model contains 220,000 mesh elements, with a mesh quality of 1, meeting the simulation requirements. For the fluid boundary conditions, the inlet and outlet are set as pressure inlet and pressure outlet, respectively. The simulation adopts the standard k–ε turbulence model, and the solver is configured as a pressure-based steady-state absolute velocity solver. The wall surfaces of the model are set to no-slip conditions. The time step is set to 0.01 s, with a total of 1000 steps, until stable flow field data is obtained.

2.3.3. DEM Simulation Model

This study uses Shuyounuo 81 as the research object. In EDEM 2023.1, two particle models were established: one model of rice grains and one of panicles. Based on the actual dimensions of the grains and heads, the grains are simplified as elliptical five-sphere clusters with a major axis of 8.0 mm and a minor axis of 3.0 mm [25,26]. The panicle consists of a 12 cm main panicle and a 6 cm stem.

The main panicle comprises a main panicle axis, 10 primary branches, and 80 grains. The main panicle axis is modeled using 30 particles with a diameter of 2 mm, and each primary branch is composed of 40 particles with a diameter of 1 mm. The discrete element models of the rice grains and panicles are shown in Figure 6a,b, respectively.

The 3D model of the header created in SolidWorks [27] (v.2022, Dassault Systèmes, Vélizy-Villacoublay, France) was converted to .igs format and imported into EDEM. We set up a 150 mm × 960 mm particle factory under the header auger blades, generating 1.51 × 10⁴ grains and 10 panicles based on prior experimental statistics, with a total weight of 533 g. Grains and panicles are generated in a particle factory with no initial velocity and fall only by gravity. Rayleigh time step 9.74 × 10⁻⁶ s in the software, and the save time is set to 0.01 s, The Hertz–Mindlin (No Slip) model was adopted as the contact model, the simulation time is 3 s.

The computational domain was set, and the number of grains blown off the header was recorded in the EDEM Solve Report. The self-cleaning rate was defined as: Self-cleaning rate Mass of blown-off particles/Mass of blown-off particles %. The simulation parameters are listed in Table 3 and Table 4.

2.3.4. CFD-DEM Coupled Single-Factor Experiment Method

To further determine the influence of various factors on the self-cleaning rate, single-factor experiments were conducted. The influencing factors considered were inlet air pressure, incident airflow angle [28], and nozzle outlet height, with the self-cleaning rate as the evaluation criterion. The factors and levels for the single-factor experiments are shown in Table 5.

2.3.5. Optimization Simulation Experiment Method

To explore the optimal combination of inlet air pressure, incident airflow angle, and nozzle outlet height on the self-cleaning rate, a three-factor, three-level central composite experiment was conducted. The experimental factors included inlet air pressure, incident airflow angle, and nozzle outlet height, with the self-cleaning rate Y as the performance evaluation indicator. The factor coding for the experiments is shown in Table 6, where X₁, X₂, X₃ represent the coded values of the factors.

2.4. One-Click Cleaning Control System

2.4.1. Control System Hardware Design

The “One-Click Cleaning” control system uses the STM32F103C8T6 microcontroller, Shanghai, China, as the control core. The hardware composition is shown in Figure 7 and primarily includes the STM32 microcontroller minimum system board, 12 V relays, an electric push rod, a 5 V step-down module, a 12 V lead–acid battery, a 12 V switching power supply, a one-click cleaning button, solenoid valves, and a DC motor.

The power supply voltages required by the components in the control system are 12 V and 5 V. 12 V voltage regulator module is connected in parallel with the lead–acid battery to power the relays, DC motor, and other components. The STM32 microcontroller receives its operating voltage through a 5 V transformer connected to the 12 V lead–acid battery. The extension and retraction of the electric push rod, as well as the forward and reverse rotation of the DC motor, can be controlled by switching the contacts of a pair (two) of relays. The air compressor operates at a voltage of 12 V with a power of 550 W. Its power supply line is connected in parallel with the harvester’s onboard battery. During the cleaning process, the engine must maintain a high RPM to ensure that the air compressor operates at full power. In order to provide a deeper understanding of the hardware circuit structure of the control system, the circuit diagram of the control body is shown in Figure 8.

2.4.2. Control System Program Design

The STM32 microcontroller serves as the core of the “One-Click Cleaning” system, primarily responsible for the logical control of the cleaning command. The STM32F103C8T6 microcontroller, used as the main chip, is a 32-bit microcontroller (MCU) based on the ARM Cortex-M3 core. It features 64 KB Flash and 20 KB SRAM, with a maximum clock frequency of 72 MHz, 48 I/O ports, and three 16-bit timers [29,30]. Although it does not directly provide standard 32-bit timers, it includes a 24-bit SysTick decrement counter, which can simulate 32-bit timer functionality in certain cases. This controller meets the hardware control requirements outlined above.

The overall control flow of the system is illustrated in Figure 9. Upon powering on, the system initializes all parameters. During the cleaning process, pressing the “One-Click Cleaning” button on the control box initiates the sequence. The electric push rod opens the aggregate bin lid, and the telescopic air duct extends the nozzle. The microcontroller then sends a high-level signal through the GPIO pin to control the corresponding solenoid valve, allowing it to open and activate the nozzle. After the solenoid valve has remained open for the set duration, it closes, stopping the airflow through the nozzle. The microcontroller subsequently sends a low-level signal through the GPIO pin to shut off the nozzle, retract it to its original position, and close the aggregate bin lid, completing the cleaning process.

2.5. Field Validation Experiment

2.5.1. Experimental Materials and Location

To evaluate the reliability and self-cleaning performance of the automatic cleaning header, field experiments were conducted at Yuankang Agricultural Cooperative, Xinlong Town, Dongfang City, Hainan Province, China (latitude 18° N, longitude 108° E). The test rice variety was Shuyounuo 81, with a yield of 180 kg per mu and a moisture content of 23% at harvest. The 1000-grain weight of the rice was 32 g. Before the experiment, the male parent plants were removed from the seed production field, and impurity removal was performed on the female parent plants. The experimental field was a drained dry field. The harvest experiment utilized an LDS-1G moisture meter, a high-precision electronic scale, and a 4LZ-7G2A (G4) Weifang China crawler-type full-feed harvester. All hardware circuits were constructed, and the header was reprocessed for the field experiments, as shown in Figure 9.

2.5.2. Experimental Method

After completing the harvesting operation, the machine was stopped, and the header was lowered to its lowest position before cleaning. First, manual cleaning was performed to collect the residue from the bottom of the header, and the collected residue was weighed to obtain the initial weight w1. The collected residue was then returned to the header, and the transverse conveying auger was activated to spread the residue as evenly as possible. The “One-Click Cleaning” button in the cabin was pressed to initiate the cleaning process. After cleaning, the residue collected in the aggregate bin was weighed again to obtain the weight w2. The self-cleaning rate (z) was calculated using the following formula:

(4)$z={\displaystyle \frac{w2}{w1}}\mathrm{\%}$

The experiment was conducted five times, and the average values were calculated to ensure the accuracy of the results (Figure 10).

3. Results and Discussion

3.1. Nozzle Optimization Simulation Results

3.1.1. Single-Factor Simulation Results for the Nozzle

The single-factor simulation results for the nozzle are shown in Figure 11. Figure 11a shows the effect of the expansion section length on the nozzle angle. The expansion section distance allows for airflow diffusion, which not only reduces turbulence and vortices at the nozzle outlet but also helps form a wider spray width. The results indicate that as the expansion section length increases, the nozzle spray angle gradually decreases, showing a significant effect of the expansion section length on the spray width. Figure 11b shows the effect curve of the contraction section length on the spray width. Contraction in the throat enables the airflow to achieve a relatively steady velocity distribution. Simulation results indicate that for contraction section lengths of 6 mm, 8 mm, and 10 mm, the corresponding results are spray angles of 55.75°, 48.42°, and 58.12°, respectively, showing significant changes in spray angle with contraction section length. Figure 11c shows the effect of throat diameter on the spray angle. Larger throat diameters weaken the airflow acceleration effect, resulting in reduced spray angles. Simulation results show that for diameters in the range of 8–12 mm, the spray angles are 64.92°, 51.64°, and 45.36°, respectively, showing a significant effect of throat diameter on spray angle. Figure 11d shows the effect of inlet air pressure on the spray angle. Simulation results indicate that under different inlet air pressures, the spray angle varies between 59.21° and 62.46°. When the nozzle shape is fixed, inlet air pressure only affects the gas velocity and has little impact on the spray angle. Figure 11e shows the effect of outlet height on the spray angle. Simulation results indicate that when the outlet height is within the range of 2–3 mm, the spray angle remains stable at 52.40°, 51.52°, and 51.40°, respectively. Outlet height has a greater effect on air consumption but little effect on the spray angle.

3.1.2. Nozzle Orthogonal Experiment Simulation Results

The simulation results of the nozzle orthogonal experiment are shown in Table 7.

From the range results of the nozzle spray width, it can be observed that the range trend R of the nozzle spray width indicates the order of factors influencing the spray width as B > C > A, which corresponds to contraction section distance, expansion section distance, and throat diameter, respectively. By comparing the K values (the sum of experimental indicators at each level for each factor in the table), the optimal level combination is A1B2C3. Thus, when the expansion section distance is 10 mm, the throat diameter is 8 mm, and the contraction section distance is 8 mm, the nozzle achieves the largest jet angle under the same pressure conditions. The spray width range of a single nozzle is calculated as follows (Figure 12):

(5)$\mathit{tan}\alpha ={\displaystyle \frac{D}{2L}}$

(6)$D=2\cdot L\cdot \mathit{tan}\alpha$

By substituting the maximum angle α into Equation (5), the spray width D is calculated to be 50.3 cm. The header width of the seed production harvester is 2.2 m. To ensure that the entire header is covered by airflow, five nozzles are required.

The three-dimensional header model was imported into the Fluent module of ANSYS 19.0. After completing the fluid domain extraction, the mesh at the connection points between the nozzles and the header was locally refined. A mesh quality check was then performed to ensure that the mesh met the analysis requirements. In the solver, the realizable k–ε turbulence model (standard wall functions, SWF) was selected, with air as the fluid. The pressure inlet was set to 0.2 Mpa, and 1500 iterations were performed. The velocity contour along the incident plane was obtained, as shown in Figure 13. The figure indicates that when five nozzles are used, the airflow can cover the entire header surface, providing a basis for subsequent cleaning experiments.

3.2. Structural Analysis of Header Cleaning CFD-DEM Coupling Simulation

Airflow velocity has a direct impact on the self-cleaning effect. The airflow vector diagram within the header during cleaning is shown in Figure 14a. The figure indicates that airflow velocity is highest near the curved bottom plate, while it decreases as the distance from the nozzle increases. This is because most of the airflow collides with the curved bottom plate and disperses. During cleaning, the incident airflow should have an appropriate angle to minimize airflow dissipation and prevent a low airflow velocity that could reduce the self-cleaning rate. The particle motion trajectory is shown in Figure 14b. Under the impact of high-speed airflow, grains and stems fall from the header with an initial velocity. However, the curvature of the header bottom plate and the presence of the transverse conveyor auger create obstructions to the airflow. Excessively large or small incident airflow angles lead to airflow dissipation, which necessitates further experimental studies on the factors affecting the header flow field.

3.2.1. CFD-DEM Single-Factor Simulation Experiment Results

In the experiment evaluating the effect of inlet air pressure on cleaning, the pre-experiment determined the airflow incident angle to be 118° and the nozzle outlet height to be 2 mm. Simulations were conducted with inlet air pressures of 0.2 Mpa, 0.3 Mpa, 0.4 Mpa, 0.5 Mpa, and 0.6 Mpa. As shown in Figure 15a, as the inlet air pressure increases, the airflow velocity through the nozzle rises, enhancing the airflow’s ability to blow away residue from the header. The self-cleaning performance improves continuously, reaching 99.2% at an inlet air pressure of 0.6 Mpa, meeting the requirements for cleaning. Higher inlet air pressures, however, lead to increased energy consumption, making 0.6 Mpa the most suitable pressure for cleaning. In the experiment assessing the effect of incident airflow angle on cleaning, the conditions were set with an inlet air pressure of 0.2 Mpa and a nozzle outlet height of 2 mm. Simulations were conducted with incident airflow angles of 116°, 117°, 118°, 119°, and 120°. As shown in Figure 15b, the self-cleaning rate initially increases and then decreases as the incident airflow angle rises. When the incident airflow angle is less than 116°, the airflow is dissipated by the obstruction from the transverse conveyor blades, reducing its blowing capacity and resulting in a lower self-cleaning rate. Conversely, when the incident angle exceeds 120°, the airflow is dispersed by the curved bottom plate of the header, negatively affecting the self-cleaning performance. Therefore, an incident airflow angle of 118° is most suitable for cleaning. In the experiment evaluating the effect of nozzle outlet height on self-cleaning performance, the conditions were set with an inlet air pressure of 0.2 Mpa and an incident airflow angle of 118°. Simulations were conducted with nozzle outlet heights of 2.0 mm, 2.2 mm, 2.4 mm, 2.6 mm, and 2.8 mm. As shown in Figure 15c, the self-cleaning rate increases as the nozzle outlet height increases due to greater airflow volume per unit of time. At an outlet height of 2.6 mm, the self-cleaning effect is optimal. However, further increases in outlet height reduce airflow velocity, lowering the self-cleaning performance. Therefore, a nozzle outlet height of 2.6 mm is most suitable for cleaning.

3.2.2. CFD-DEM Optimized Simulation Experiment Results

The results of the CFD-DEM optimized simulation experiments are shown in Table 8.

3.2.3. Regression Equation and Significance Test

Using Design-Expert 8.0 software and the Box–Behnken central composite design principle [31], the experiment was designed, and a quadratic polynomial regression model for the self-cleaning rate was obtained as follows:

y = 98.56 + 1.76A + 1.13B + 1.08C + 0.94AB + 0.79AC − 0.77BC − 0.86A² − 3.68B² − 3.74C²

According to Table 9, the p-value of the regression model is less than 0.0001, indicating that the model is highly significant. The p-value of the lack-of-fit term is greater than 0.05, suggesting that the lack of fit is not significant and that the regression model has a high degree of fit. From the p-values of the factors inlet air pressure, incident airflow angle, and nozzle outlet height, it can be determined that these three factors have a highly significant effect on the self-cleaning rate. The order of influence on the self-cleaning rate from largest to smallest is inlet air pressure, incident airflow angle, and nozzle outlet height.

The determination coefficient R² of the self-cleaning rate model is 0.9845, indicating that the model can explain 98.45% of the experimental results. The adjusted determination coefficient is close to one, demonstrating that the regression model has high reliability.

3.2.4. Response Surface of Interaction Effects on Self-Cleaning Rate

To visually analyze the interaction effects of experimental factors on the self-cleaning rate, the response surface plots derived from the regression model are shown in Figure 16.

Figure 16a shows the interaction effects of inlet air pressure and incident airflow angle on the self-cleaning rate when the nozzle outlet height is 2.6 mm. When the inlet air pressure is constant, the self-cleaning rate initially increases and then decreases as the incident airflow angle increases. The reason is that when the incident angle of the airflow direction is 117°, part of the airflow hits the transverse conveying auger and causes airflow loss, and with the increase of the incident angle, the airflow gradually moves towards the arc bottom plate. When the angle of the incident airflow continues to increase to 119°, part of the airflow hits the arc floor and the airflow loss increases. As the inlet air pressure increases, the rate of decline in the self-cleaning rate caused by an increasing incident angle slows down. When the incident airflow angle is constant, the self-cleaning rate increases with increasing inlet air pressure; this is because greater air pressure leads to greater air velocity, and the initial velocity of grain movement increases. The maximum self-cleaning rate occurs when the inlet air pressure is 0.5–0.6 Mpa, and the incident airflow angle is 117.5–118.5°.

Figure 16b shows the interaction effects of inlet air pressure and nozzle outlet height on the self-cleaning rate when the incident airflow angle is 118°. When the nozzle outlet height is constant, the self-cleaning rate increases with increasing inlet air pressure. When the inlet air pressure is constant, the self-cleaning rate initially increases and then decreases with increasing nozzle outlet height.The reason for this is that when the nozzle outlet height is 2.4 mm, the flow rate decreases due to the decrease in the flow area when the air flow passes through the nozzle. When the height of the nozzle outlet continues to increase to 2.8 mm, the acceleration effect of the nozzle on the airflow decreases, the airflow velocity of the nozzle decreases, and the self-cleaning rate decreases. The maximum self-cleaning rate occurs when the inlet air pressure is 0.5–0.6 Mpa and the nozzle outlet height is 2.5–2.7 mm.

Figure 16c shows the interaction effects of incident airflow angle and nozzle outlet height on the self-cleaning rate when the inlet air pressure is 0.5 Mpa. When the nozzle outlet height is constant, the self-cleaning rate initially increases and then decreases as the incident airflow angle increases. When the incident airflow angle is constant, the self-cleaning rate initially increases and then decreases with increasing nozzle outlet height. The maximum self-cleaning rate occurs when the incident airflow angle is 117.5–118.5° and the nozzle outlet height is 2.5–3.0 mm.

3.2.5. Parameter Optimization

To achieve the optimal combination of experimental factors for airflow cleaning performance, with the maximum self-cleaning rate as the ultimate optimization goal, the established quadratic regression model was subjected to multi-factor optimization. The objective function and constraint conditions are as follows:

(7)$\left\{\begin{array}{l}max y\hspace{0.25em}\\ \\ \mathrm{s}.\mathrm{t}.\left\{\begin{array}{l}0.2 \mathrm{Mpa}\le \mathrm{X}1\le 0.6 \mathrm{Mpa}\\ 116\le \mathrm{X}2\le 120°\\ 2 \mathrm{m}\mathrm{m}\le \mathrm{X}3\le 3 \mathrm{m}\mathrm{m}\end{array}\right.\end{array}\right.$

Using the Optimization module of the Design-Expert software, the solution was obtained: when the inlet air pressure is 0.6 Mpa, the incident airflow angle is 118.25°, and the nozzle outlet height is 2.64 mm, the self-cleaning rate reaches its maximum value of 99.63%.

3.3. Cleaning Performance Indicators and Field Test Validation

According to the national standard GB4404.1-2008 Seeds of Food Crops—Part 1: Cereal Crops, the purity of hybrid rice seeds for field use must not be lower than 95.0% [32]. Therefore, the rice seed production harvester must completely clean the header with no dead zones before harvesting different rice varieties to prevent cross-contamination. Under the optimal parameter combination, namely, inlet air pressure of 0.6 Mpa, incident airflow angle of 118.25°, and nozzle outlet height of 2.64 mm, the self-cleaning rate of the rice seed production harvester reached 97.68%. The experimental results are shown in Figure 16, meeting the performance requirements of the seed production harvester.

It is worth noting that the actual self-cleaning rate was slightly lower than the simulation results. This discrepancy can be attributed to the challenges in achieving the exact precision of the nozzle and telescopic tube during the manufacturing process, as well as airflow pressure loss within the pipes during the cleaning operation, which reduces the outlet pressure (Figure 17).

4. Conclusions

(1) This study designed a concealed automatic cleaning header to address the issue of residue accumulation in the header of rice seed production harvesters. The header structure was redesigned, featuring a concealed nozzle that remains closed during harvesting operations and opens during cleaning, avoiding interference with harvesting operations while protecting the nozzle. Additionally, an aggregate bin was designed to prevent seed source waste. A “one-click cleaning” button was developed for the driver’s cabin, enabling intelligent cleaning by controlling the aggregate bin lid, cleaning nozzle, and solenoid valves via an STM32 microcontroller. At the same time, the cleaning machine solution has good versatility. All nozzles perform the same cleaning action under the control of the control system; therefore, the number of nozzles only needs to be adjusted according to the width of the header to ensure that the airflow can fully cover the entire header and can be adapted to different widths of the header, and the versatility of the cleaning function can reduce the cost of use for farmers.

(2) The nozzle was identified as the critical component for cleaning. The nozzle structure is hidden. Each nozzle corresponds to a reserved opening on the rear plate of the header, which increases the strength of the rear plate and increases the complexity of the control system, and its structure was designed accordingly. Through single-factor experiments and orthogonal combination experiments, simulation tests were conducted to analyze the parameters affecting the nozzle spray width. The optimal nozzle structure was determined, with an expansion section distance of 10 mm, a throat diameter of 8 mm, and a contraction section distance of 8 mm, achieving the widest spray coverage. The number of nozzles required at this point is five.

(3) Based on gas–solid coupling simulations, a controlled variable method was employed to simulate multiple combinations of inlet air pressure, incident airflow angle, and nozzle outlet height. The optimal self-cleaning performance was achieved when the inlet air pressure was 0.6 Mpa, the incident airflow angle was 118.25°, and the nozzle outlet height was 2.64 mm, with a self-cleaning rate of 99.63%. Field experiments conducted under these optimal parameters resulted in a self-cleaning rate of 97.68%. The simulation results are in good agreement with the actual field test results, which further proves the effectiveness of the Fluent–EDEM coupling simulation method in the design of the cleaning device. The single cleaning time of 10 s meets the requirements for rapid and effective self-cleaning of the header in rice seed production harvesters.

Author Contributions

Conceptualization, Y.Q. and L.C.; methodology, Y.Q.; software, L.C.; validation, P.W. and D.C.; formal analysis, Y.Q.; investigation, D.C., W.S. and R.Y.; resources R.Y.; data curation, D.C. and W.S.; writing—original draft preparation, Y.Q. and L.C.; writing—review and editing, Y.Q. and L.C.; visualization, Y.Q. and D.C.; supervision, R.Y.; project administration, Y.Q.; funding acquisition, Y.Q. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement

The data presented in this study are available on demand from the second author at [email protected].

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Figure 1. Schematic diagram of the hidden self-cleaning header device. 1. Dial wheel. 2. Aggregate bin cover. 3. Cutter. 4. Divider. 5. Aggregate bins. 6. Nozzle telescopic tube. 7. Auger with spiral conveyor.

Figure 2. Schematic diagram of the key components of the header. 1. Electric actuators. 2. Aggregate bin floor. 3. Aggregate bin cover/header bottom plate. 4. Aggregate bin silo. 5. Rack-and-pinion drive. 6. Nozzle inner tube. 7. Nozzle. 8. DC motor.

Figure 3. Schematic diagram of the working diagram and nozzle structure of the air flow cleaner. 1. Auger. 2. Nozzle. 3. Curved base plate. 4. Material. 5. Aggregate bin cover.

Figure 4. Nozzle airflow contour fitting curve.

Figure 5. Fluent pre-processing.

Figure 6. Discrete element model of grain and panicle.

Figure 7. Diagram of the hardware composition of the control system.

Figure 8. Schematic diagram of the control system circuit.

Figure 9. Control system flow diagram.

Figure 10. Field validation experiment.

Figure 11. Nozzle single-factor simulation results.

Figure 12. Nozzle spray width.

Figure 13. Simulated contour of header nozzle.

Figure 14. Diagram of the results of the coupled simulation.

Figure 15. Simulation results of different single factors on self-cleaning rate.

Figure 16. Effect of interaction of factors on cleanliness. (a) The interaction effects of inlet air pressure and incident airflow angle on the self-cleaning rate when the nozzle outlet height is 2.6 mm. (b) The interaction effects of inlet air pressure and nozzle outlet height on the self-cleaning rate when the incident airflow angle is 118°. (c) The interaction effects of incident airflow angle and nozzle outlet height on the self-cleaning rate when the inlet air pressure is 0.5 Mpa.

Figure 17. Renderings of the cleaning machine.

Table 1

Single-factor experiments level table.

Test Group	Factors
	Expansion Section Distance/mm	Contraction Section Distance/mm	Throat Contraction Diameter/mm	Inlet Air Pressure/Mpa	Nozzle Outlet Height/mm
1	10 15 20 25 30	8	10	0.4	2.5
2	15	6 8 10 12 14	10	0.4	2.5
3	15	8	16 14 12 10 8	0.4	2.5
4	15	8	10	0.1 0.2 0.3 0.4 0.5	2.5
5	15	8	10	0.4	1.5 2 2.5 3 3.5

Table 2

Table of orthogonal experiment factor levels.

Levels	A	B	C
	d1/mm	d2/mm	d3/mm
1	10	6	12
2	15	8	10
3	20	10	8

Table 3

Material parameters.

Simulated Materia	Poisson’s Ratio	Elastic Modulus	Density
Rice Grain	0.28	375	1350
Rice Stem	0.4	2.8	215
Steel	0.3	2.06 × 105	7800

Table 4

Contact parameters between materials.

Contact Parameter	Restitution Coefficient	Static Friction Coefficient	Rolling Friction Coefficient
Grain–Grain	0.5	0.425	0.01
Grain–Stem	0.2	0.8	0.01
Grain–Steel Plate	0.5	0.58	0.01
Stem–Stem	0.2	0.9	0.01
Stem–Steel	0.3	2.06 × 105	7800

Table 5

Factors and levels of single-factor experiment.

Experimental Group	Inlet AirPressure/Mpa	Incident Angle/°	Nozzle Outlet Height/mm
1	0.2 0.3 0.4 0.5 0.6	118	2.0
2	0.2	116 117 118 119 120	2.0
3	0.2	118	2.0 2.2 2.4 2.6 2.8

Table 6

Experimental factor coding.

Code	X1	X2	X3
	Inlet Air Pressure/Mpa	Incident Airflow Angle/°	Nozzle Outlet Height/mm
−1	0.4	117	2.4
0	0.5	118	2.6
1	0.6	119	2.8

Table 7

Orthogonal experimental results.

Test Number	A	B	C	Jet Angle/°
1	10	6	12	55.74
2	15	8	12	48.60
3	20	10	12	25.34
4	15	6	10	38.17
5	20	8	10	39.54
6	10	10	10	30.54
7	20	6	8	42.44
8	10	8	8	77.96
9	15	10	8	47.26
K1	164.24	136.35	129.68
K2	134.03	166.10	108.25
K3	107.32	103.14	167.66
K1	54.75	45.45	43.23
K2	44.68	55.37	36.08
K3	35.77	34.38	55.89
R	18.97	20.99	19.80

Table 8

Experimental protocols and results.

Test Number	X1	X2	X3	Self-Cleaning Rate/(%)Y
1	−1	−1	0	91.92
2	1	−1	0	93.33
3	−1	1	0	92.83
4	1	1	0	97.99
5	−1	0	−1	91.94
6	1	0	−1	94.11
7	−1	0	1	92.24
8	1	0	1	97.55
9	0	−1	−1	88.27
10	0	1	−1	91.56
11	0	−1	1	92.25
12	0	1	1	92.45
13	0	0	0	98.45
14	0	0	0	98.63
15	0	0	0	97.62
16	0	0	0	99.63
17	0	0	0	98.45

Table 9

Significance test results.

Source	Sum of Squares	df	Mean Square	F-Value	p-Value	Significant
Model	181.85	9	20.21	49.43	<0.0001	significant
A	24.68	1	24.68	60.36	0.0001
B	10.26	1	10.26	25.10	0.0015
C	9.27	1	9.27	22.67	0.0021
AB	3.52	1	3.52	8.60	0.0219
AC	2.46	1	2.46	6.03	0.0437
BC	2.39	1	2.39	5.84	0.0463
A2	3.08	1	3.08	7.54	0.0287
B2	57.11	1	57.11	139.72	<0.0001
C2	58.91	1	58.91	144.11	<0.0001
Residual	2.86	7	0.4088
Lack of fit	0.8039	3	0.2680	0.5210	0.6905	not significant
Pure error	2.06	4	0.5144
Cor total	184.71	16

Note: p ≤ 0.01, highly significant; p ≤ 0.05, significant; p > 0.1 not significant.

References

1. Chen, Q.; Han, Z.; Cui, J.; Wang, G.; Qiao, X.; Zhang, Z.; Han, K.; Gan, B. Development Status and Trend Current Situation of Self-propelled Combine Harvester. J. Agric. Sci. Technol.; 2015; 17, pp. 109-114. [DOI: https://dx.doi.org/10.13304/j.nykjdb.2015.042]

2. Luo, X.; Zhou, Y. Potential Food Security Risks and Countermeasures under the Background of Seed Industry Innovation Based on Industry 4.0. Mob. Inf. Syst.; 2022; 2022, 9905894. [DOI: https://dx.doi.org/10.1155/2022/9905894]

3. Shang, S.; Wu, X.; Yang, R.; Li, G.; Yang, X.; Chen, D. Research Status and Prospect of Plot-sowing Equipment and Technology. Trans. Chin. Soc. Agric. Mach.; 2021; 52, pp. 1-20. [DOI: https://dx.doi.org/10.6041/j.issn.1000-1298.2021.02.001]

4. Dai, F.; Han, Z.; Zhang, F.; Gao, A.; Sun, B. Experiment on longitudinal axial conical cylinder threshing separation unit for plot breeding. Agric. Res. Arid Areas; 2012; 30, pp. 274-278.

5. Zhu, M.; Chen, H.; Li, Y. Investigation and development analysis of seed industry mechanization in China. Trans. Chin. Soc. Agric. Eng. (Trans. CSAE); 2015; 31, pp. 1-7. [DOI: https://dx.doi.org/10.11975/j.issn.1002-6819.2015.14.001]

6. Zhao, C.; Han, Z.; Cao, Z. Development of the Hand Supporting Cabine Harvesterof Seeds with Air Suction Stripping and C leaning on Breeding Plot. J. Chin. Agric. Mech.; 2010; 4, pp. 64-67. [DOI: https://dx.doi.org/10.3901/JME.2010.20.064]

7. Dai, F.; Song, X.F.; Zhao, W.Y.; Han, Z.S.; Zhang, F.W.; Zhang, S.L. Motion simulation and test on threshed grains in tapered threshing and transmission device for plot wheat breeding based on CFD-DEM. Int. J. Agric. Biol. Eng.; 2019; 12, pp. 66-73. [DOI: https://dx.doi.org/10.25165/j.ijabe.20191201.4458]

8. Wang, J.; Wang, D.; Shang, S.; Wang, Y.; Kuai, J. Development and experiment on 4LZZ-1.0 type plot grain combine. Trans. Chin. Soc. Agric. Eng. (Trans. CSAE); 2016; 32, pp. 19-25. [DOI: https://dx.doi.org/10.11975/j.issn.1002-6819.2016.18.003]

9. Yao, W. Research and Design of Key Technology and Components of Wheat for Breeding Pots Harvester. Master’s Thesis; Shandong Agricultural University: Taian, China, 2017.

10. Xia, S.; Li, Y.; Ji, K.; Du, T.; Wang, F. Simulation Analysis of Air Flow Seed Cleaning for Seed Combine Harvester Based on CFD-DEM Simulation Analysis of Air Flow Seed Cleaning for Seed Combine Harvester Based on CFD-DEM. J. Agric. Mech. Res.; 2024; 47, pp. 1-8. [DOI: https://dx.doi.org/10.13427/j.cnki.njyi.20240703.002]

11. Ding, B.; Liang, Z.; Qi, Y.; Ye, Z.; Zhou, J. Improving Cleaning Performance of Rice Combine Harvesters by DEM-CFD Coupling Technology. Agric. Basel; 2022; 12, 1457. [DOI: https://dx.doi.org/10.3390/agriculture12091457]

12. Qu, Z.; Lu, Q.; Shao, H.; Le, J.; Wang, X.; Zhao, H.; Wang, W. Design and Test of a Grain Cleaning Loss Monitoring Device for Wheat Combine Harvester. Agriculture; 2024; 14, 671. [DOI: https://dx.doi.org/10.3390/agriculture14050671]

13. Wang, J.; Liao, M.; Xia, H.; Chen, R.; Li, J.; Li, J.; Yang, J. Study on a Novel Reseeding Device of a Precision Potato Planter. Agric. Basel.; 2024; 14, 1824. [DOI: https://dx.doi.org/10.3390/agriculture14101824]

14. Zha, X.; Zhang, G.; Han, Y.; Fu, J.; Zhou, Y. Design and Optimization of Centralized Pneumatic Fertilizer Transporting System for Deep Fertilization Device in Paddy Field. Int. J. Fluid Mach. Syst.; 2021; 14, pp. 309-321. [DOI: https://dx.doi.org/10.5293/IJFMS.2021.14.3.309]

15. Ding, H. Design and Research of Self-Propelled Variable-Width High-Pressure Water Surface Cleaning Machine. Master’s Thesis; Chang’an University: Xi’an, China, 2020.

16. Zhang, H.; Wang, S.; Zheng, Y.; Wang, W.; Li, W.; Wang, P.; Ren, G.; Ma, J. FPGA implementation of image edge detection based on four algorithms of GAUSS-filter, Sobel, NMS and OTSU. Chin. J. Liq. Cryst. Disp.; 2020; 35, pp. 250-261. [DOI: https://dx.doi.org/10.3788/YJYXS20203503.0250]

17. Fang, Y.; Xie, W.; Liu, Y.; He, L. Optimization of Air Jet Nozzle Structure and Experimental Research on Arch Breaking. Mach. Hydraul.; 2019; 47, pp. 49-53. [DOI: https://dx.doi.org/10.3969/j.issn.1001-3881.2019.07.011]

18. Xin, N. Simulation Analysis of Working Process of Air-Blowing Seed-Metering Device Based on Coupled EDEM-FLUENT. Master’s Thesis; Jilin University: Changchun, China, 2013.

19. Xin, Q. Simulation and Test of Potato Seedling Crushing and Throwing Device Based on Fluent-EDEM. Master’s Thesis; Shandong Agricultural University: Taian, China, 2020.

20. Xia, X. Research on Numerical Modeling Method of Cotton Picking Device and Flow Characteristics of Seed Cotton Complex Coupled Airflow Field. Master’s Thesis; Xinjiang University: Urumqi, China, 2021.

21. Xu, L.; Li, Y.; Chai, X.; Wang, G.; Liang, Z.; Li, Y.; Li, B. Numerical simulation of gas–solid two-phase flow to predict the cleaning performance of rice combine harvesters. Biosyst. Eng.; 2020; 190, pp. 11-24. [DOI: https://dx.doi.org/10.1016/j.biosystemseng.2019.11.014]

22. Wang, L.; Yang, H.; Wang, Q.; Lu, C.; Wang, C. Design and Experiment of Air-assisted Linear Seeding Device for Wheat. Trans. Chin. Soc. Agric. Mach.; 2024; 55, pp. 121–132+158. [DOI: https://dx.doi.org/10.6041/j.issn.1000-1298.2024.06.011]

23. Pan, J. The Design and Experimental Research of Air Blow Type Ground Jujube Picking Device. Master’s Thesis; Shihezi University: Shihezi, China, 2018.

24. Chen, S.; Wang, Z.; Wu, W.; Liu, C. Simulating the circulation and longitudinal ventilation of corn storage silos using EDEM-Fluent. Trans. Chin. Soc. Agric. Eng.; 2024; 40, pp. 40-49. [DOI: https://dx.doi.org/10.11975/j.issn.1002-6819.202307126]

25. Su, Z.; Li, Y.; Dong, Y.; Tang, Z.; Liang, Z. Simulation of rice threshing performance with concentric and non-concentric threshing gaps. Biosyst. Eng.; 2020; 197, pp. 270-284. [DOI: https://dx.doi.org/10.1016/j.biosystemseng.2020.05.020]

26. Liang, Z.; Li, J.; Liang, J.; Shao, Y.; Zhou, T.; Si, Z.; Li, Y. Investigation into Experimental and DEM Simulation of Guide Blade Optimum Arrangement in Multi-Rotor Combine Harvesters. Agriculture; 2022; 12, 435. [DOI: https://dx.doi.org/10.3390/agriculture12030435]

27. Liu, Y.; Li, Y.; Chen, L.; Zhang, T.; Liang, Z.; Huang, M.; Su, Z. Study on Performance of Concentric Threshing Device with Multi-Threshing Gaps for Rice Combines. Agriculture; 2021; 11, 1000. [DOI: https://dx.doi.org/10.3390/agriculture11101000]

28. Li, X. Study of High-Pressure Jet Characteristic of Fan-Shaped Nozzle. Master’s Thesis; Southwest Jiaotong University: Chengdu, China, 2014.

29. Yang, S.; Wang, X.; Gao, Y.; Zhao, X.; Dou, H.; Zhao, C. Design and Experiment of Motor Driving Bus Control System for Corn Vacuum Seed Meter. Trans. Chin. Soc. Agric. Mach.; 2019; 50, pp. 57-67. [DOI: https://dx.doi.org/10.6041/j.issn.1000-1298.2019.02.007]

30. Zhang, S.; Wei, X.; Deng, Y.; Zang, C.; Ji, X.; Wang, A. Design and Experiments of the Whole Field Path Tracking Algorithm for a Track-Based Harvester. Trans. Chin. Soc. Agric. Eng.; 2023; 39, pp. 36-45. [DOI: https://dx.doi.org/10.11975/j.issn.1002-6819.202303162]

31. Li, X.; Wei, Z.; Su, G.; Wang, X.; Zhang, X.; Wang, X.; Chang, X.; Cheng, X.; Jin, C. Parameter Optimization and Experiment of Air-screen Potato-stone Cleaning Device Based on Gas-Solid Coupling. Trans. Chin. Soc. Agric. Mach.; 2023; 55, pp. 96-108. [DOI: https://dx.doi.org/10.6041/j.issn.1000-1298.2024.01.009]

32. Tang, Z.; Liu, S.; Zhou, F.; Li, T.; Wang, J.; Li, C. Design and experiment of ear harvester for seed corn. Trans. Chin. Soc. Agric. Mach.; 2021; 52, pp. 102-112. [DOI: https://dx.doi.org/10.6041/j.issn.1000-1298.2021.07.010]