Abstract:Discrete element method (DEM) has emerged as an important approach for investigating the operational mechanisms of agricultural equipment and optimizing their operation and structure parameters. By establishing reliable numerical models, DEM enables the simulation of interactions between agricultural equipment and agricultural materials. In recent years, scholars worldwide have developed a series of agricultural material models using DEM, achieving significant progress in simulating equipment-material interactions across various agricultural processes. The DEM modeling technologies and key parameter calibration methods for typical agricultural materials were systematically reviewed, such as soil, seeds, crop straw, fertilizers, etc. It summarized application cases of DEM in critical agricultural operations of tillage, seeding, crop management, harvesting, and post-harvest processing and storage, and equipment design research based on DEM simulations. Current challenges were critically examined: limited model accuracy and versatility across diverse soil textures and crop varieties, computational inefficiency in large-scale field simulations, insufficient coupling application with other numerical methods such as DEM-CFD, DEM-MBD and DEM-FEM. Future research directions of the application of DEM in agricultural equipment research was proposed, including improve model accuracy and universality, enhance research on basic theory and intrinsic model, increase simulation efficiency and expand its use coupling with other methods.

Abstract:As an advanced numerical simulation method, coupled simulation technology can combine multiple independent simulation models of physical fields or physical phenomena to realize the combination of numerical simulation methods in different fields or different physical models, exchange information and interact with each other. It showed a broad application prospect in the field of agricultural engineering and provided a way for the digital and intelligent design of modern agricultural equipment. In order to systematically summarize the application status of coupling simulation technology in the field of agricultural engineering and analyze the future development trend of this technology, the basic principle, simulation process and types of coupling simulation technology and related numerical simulation technology and software were summarized. The application examples of discrete element-structural finite element coupling, discrete element-computational fluid dynamics coupling, discrete element-multibody dynamics coupling, continuum fluid-solid coupling and rigid-flexible coupling in the field of agricultural engineering were sorted out. At the same time, combined with the application status of coupling simulation technology, the problems existing in the current technology in model construction and accuracy, algorithm and computing resources, data integration and processing, simulation software design and standardization were analyzed. The application development trends of coupling simulation technology were proposed, including multi-scale modeling and intelligent verification system, high-performance algorithm and computing power, intelligent data governance and interaction mode, standardization construction, intelligent integration and scene expansion, etc., which provided reference for the further application and development of coupling simulation technology in the field of agricultural engineering.

Abstract:Wide-seedling-belt uniform sowing is an effective approach to improve wheat yield per unit area. To addresses the issues of poor uniformity and insufficient seedling belt width in existing wheat seeders for wide-seedling-belt, a pneumatic swash plate wheat wide-seedling-belt uniform sowing device was designed. Through theoretical analysis, a mechanical model of the seeding plate during the filling, carrying, and throwing processes was developed, which helped to identify and resolve the key factors influencing the performance of the seed metering device. Simulation tests and bench tests were conducted to determine the optimal structural and working parameters of the seed metering device. The results of the CFD simulation tests indicated that under the optimal structural parameters, specifically a hole diameter of 1.91mm, a cylindrical hole type, and a negative-pressure suction port diameter of 32.18mm, the flow velocity within the holes and the pressure difference across the holes were in an optimal state for seeding. The CFD-DEM coupling simulation tests further revealed that when the tilt angle of the seeding plate was set at 20°, the device exhibited an excellent population disturbance effect and minimal seed migration resistance. Furthermore, the bench tests determined that the optimal combination of working parameters was a rotational speed of 30r/min and a suction negative pressure of 4500Pa. Under these conditions, the row consistency variation coefficient (RCV), total consistency variation coefficient (TCV), and sowing uniformity variation coefficient (SUV) of each row were 2.04%, 1.45%, and 9.87%, respectively. These values met the requirements for wheat wide-seedling-belt uniform sowing. The findings can provide technical support for the design and application of wheat wide-seedling-belt uniform seeders.

Abstract:Aiming at the problems of low mechanization degree and poor stem-soil separation effect as well as severe damage during the mechanized harvesting of lily bulbs, based on the basic physical and mechanical properties of lily bulbs, a set of elastic multi-toed rubbing roller group was designed by applying bionics principles. By measuring the basic physical and mechanical parameters of lily bulbs, it was found that their average size was 83.7mm×74.6mm×49.0mm, density was 1.09g/cm3, and the rolling friction coefficient with polyurethane material was 0.42, while the static friction coefficient was 0.50. According to the requirements of lily bulb harvesting, the dynamic differential equation of lily bulbs on the elastic multi-toed rubbing roller group was established. The key factors affecting the harvesting performance of lily bulbs were found to be: the length of the rubbing toe (L), the rotational speed of the roller shaft (ω), and the number of toes per week (N). A simulation model of the interaction between the elastic multi-toed rubbing roller group, soil, and lily bulbs was established in the EDEM software. With the clear stem rate Y1 and the maximum collision force Y2 of lily bulbs as evaluation indicators, a three-factor three-level orthogonal experiment was conducted by using the Box-Behnken central composite design method. Through the construction of the objective function and constraint conditions for optimization approximation, the optimal parameter combination was determined as: L was 52mm, ω was 54r/min and N was 6. Under the optimal parameter combination, the clear stem rate Y1 of lily bulbs was 97.25% and the maximum collision force Y2 was 12N. Field experiments verified that the relative errors between the simulation values and the experimental values of the clear stem rate and the maximum collision force were 1.51% and 21.05%, respectively. Moreover, the trial-made elastic multi-toed rubbing lily harvesting device, compared with the traditional conveyor chain-type harvesting device, increased the clear stem rate by 3.3% and reduced the damage rate by 13%, demonstrating excellent harvesting performance, and meeting the relevant requirements for the quality evaluation of root and rhizome Chinese medicinal material harvesters. The research results can provide a solution for the mechanized harvesting of root and rhizome Chinese medicinal materials.

Abstract:In response to the issues of insufficient fertilizer supply and pulsating strip breakage during high-speed operation of the existing external trough-wheel fertilizer spreaders, which result in poor continuity and uniformity of fertilizer application, a high-speed air-assisted fertilizer distribution system was designed. This system featured an impeller for fertilizer supply, centrifugal fertilizer distribution, and air-flow assistance for transportation. The structural designs of the impeller feeding device, centrifugal distribution device flow deflector, and concave disk were completed. EDEM simulation experiments were conducted to verify the performance of the impeller feeding device, achieving a coefficient of variation of fertilizer supply stability below 3.2%. It was determined that the fertilizer particle velocity was the fastest when the flow deflector curve followed a friction-based brachistochrone line. The optimal structural parameters of three concave disk designs—linear, quadratic polynomial, and cubic polynomial—were determined. A bench test compared the performance of these three designs, and the linear concave disk had the most significant impact on operational effectiveness. Using the simulation-determined fertilizer feeding device, brachistochrone flow deflector, and linear concave disk, a matching experiment was conducted to determine the working parameters of the impeller speed and centrifugal disk speed. Results showed that under the condition that the variation coefficient of the consistency of fertilizer discharge volume was below 7%, when the impeller speed was 14.00~21.23r/min (corresponding to a forward speed of 9.00~10.24km/h), the centrifugal disk speed should exceed 168.2r/min;when the impeller speed was 21.23~25.00r/min (corresponding to a forward speed of 10.24~12.00km/h), the centrifugal disk speed should be below 150r/min. Field performance validation tests demonstrated that under optimal parameter matching, the variation coefficient of the stability of fertilizer discharge volume was not more than 4.11%, and the variation coefficient of the consistency of fertilizer discharge volume was not more than 6.94%. The research result can provide a reference for the design of high-speed fertilization devices for dual-purpose seeding machines used for crops such as rapeseed and wheat.

Abstract:In order to solve the problem of high grain breakage rate and unthreshed grain rate of the existing wheat plot breeding threshing machine under the condition of multi-panicle threshing, a plot breeding and threshing system based on grain-tooth combined threshing element was proposed. Combined with the structure and working principle of the breeding threshing system, the contact mechanical analysis of wheat threshing separation and the structural design of threshing components were carried out. A discrete element model of flexible hollow stems of wheat was constructed, and the effects of the forward inclination angle, ridge width and ridge height of threshing component on the threshing performance of wheat breeding were explored by taking the unthreshed rate and breakage rate of grains as evaluation indexes. The Box-Behnken analysis method was used to construct a regression model of wheat threshing loss, and the optimal structural parameters of the grain-tooth combined threshing component were determined, with front bevel angle of 56°, ridge width of 13.9mm and ridge height of 3.4mm. The prototype trial production and test verification showed that when the drum speed was 800r/min and the feeding amount was 15 ears/time, the unthreshed grain rate of wheat grains was 0.71% and grain breakage rate was 0.14%. The effectiveness and superiority of the grain-tooth combined threshing system designed in wheat multi-panicle breeding and threshing operation were verified.

Abstract:Aiming at the high moisture content wheat harvested in hilly and mountainous areas represented by Guizhou, the traditional wheat thresher can not effectively realize the cleaning separation of high moisture wheat due to insufficient power. A isolated-motor-driven cleaning technology was proposed to reduce the impurity rate and loss rate for high moisture wheat cleaning in segmented harvesting. According to the working principle of the cleaning system and aerodynamic theory, the grain movement process was analyzed and the key factors influencing the wheat cleaning performance were identified. A numerical simulation model with fluid-solid coupling was constructed by integrating computational fluid dynamics and discrete element method (CFD-DEM). Taking the impurity rate and loss rate as the evaluation indexes, the influence of the tilting angle of guide tube, cross-sectional area of suction tube and negative pressure at the suction port on the cleaning performance was investigated. The non-dominated sorting genetic algorithm (NSGA-Ⅱ) was used to perform the optimization design of key parameters matching. When the tilting angle of guide pipe was 24.0088°, the cross-sectional area of suction pipe was 179.8062cm2, and the negative pressure at the suction port was 873.6316Pa, the impurity rate and loss rate of the cleaning system were 0.92% and 0.96%, respectively. Prototype trial manufacture and experimental verification was carried out, the results showed that the cleaning impurity rate and loss rate were 1.21% and 0.82%, respectively. Compared with the traditional wheat thresher, the impurity rate was reduced by 35.98% with the loss ratio remaining no large change. The research results can provide theoretical and technical support for the research and development of high moisture wheat cleaning equipment of hilly and mountainous areas.

Abstract:In order to solve the problems of potato grading device with large structure, complex transmission and susceptibility to damage potatoes, a variable pitch potato grading device was designed, and its working principle was expounded. The key structural parameters of the potato were determined by dynamic and kinematic analysis of the potato in the grading drum. The dynamic model of the grading device was constructed by using RecurDyn software. The potato model was constructed in EDEM based on the physical characteristics of potato. The EDEM-RecurDyn coupling simulation was used to simulate the movement, constraint and contact between multiple components in the actual working process of the grading device. The amount of potato, the rotation speed of the grading drum and the length of the grading section were selected as the influencing factors, and the potato grading accuracy and grading efficiency were used as the evaluation indexes to verify the performance. Based on the above factors, the single factor and quadratic orthogonal rotation regression combination simulation test was carried out to determine the optimal structure and motion parameters of the grading device. The simulation results showed that the amount of potato was 3.68kg/s, the speed of the grading roller was 34.50r/min, and the length of the grading section was 1600mm, the optimal values of the classification accuracy and classification efficiency obtained were 91.6% and 3.0kg/s, respectively. The prototype test results showed that the average classification accuracy of the grading device was greater than 89% under the optimal solution parameters, the average classification efficiency was greater than 3.18kg/s, and the potato injury rate was less than 1%. The relative error between the simulation test and the prototype test was 2.10%, indicating that the simulation test had high accuracy and the grading device designed had good working performance, which can meet the requirements of potato grading in small and medium-sized farms.

Abstract:The first critical quality point in the wine making process is grape breakage. The grape crushing rate, juice yield rate, and the repeatability will affect the taste and flavor of wine production. The traditional centrifugal crushing machines used in winemaking often result in uneven crushing and low repeatability of crushing rates, leading to instability in the skin rupture rate and juice yield, affecting the quality of subsequent maceration and fermentation processes. To address these issues, a coaxial differential-speed centrifugal crushing method for winemaking grapes was designed, utilizing a guiding impeller to divert and direct the flow and a striking impeller to impact and crush the grapes. A mechanical model was constructed for the feeding process, flow guiding, and impact crushing to analyze and identify the key structural parameters affecting crushing performance. A model was established by using EDEM discrete element simulation to simulate the interaction between the crusher and the grapes. A three-factor, three-level response surface experiment was conducted to analyze the effects of guiding impeller speed, striking impeller speed, and feed rate on the flow uniformity and crushing stability coefficient. The optimal operating parameter combination was determined as follows: guiding impeller speed (nd) was 80r/min, striking impeller speed (nc) was 600r/min, and feed rate (wr) was 6kg/s. Performance crushing tests conducted at the factory revealed that the average crushing rate and juice yield were 96.37% and 7.05%, respectively. The variation coefficients of the crushing rate and juice yield were 1.23% and 19.76%, respectively. Compared with traditional crushers, the variation coefficient of the crushing rate was reduced by 58.3%, and that of the juice yield was reduced by 23.4%. Overall, the operational performance of the coaxial differential-speed centrifugal grape crusher met the requirements of winemaking processes. This research could improve the mechanical crushing performance of wine grapes, furthermore enhancing the quality of wine production. The research result can provide reference and guidance for the improvement of fruit crushing machinery in the relevant fruit wine brewing process.

Abstract:Simulation analysis of wheat plant cutting process based on discrete element method (DEM) is of great significance for the optimization of cutting parts of wheat harvesting machinery, while accurate measurement and calibration of mechanical parameters of wheat plant is an important prerequisite for simulation. Firstly, a wheat straw model based on the Hertz-Mindlin with Bonding model was established in the self-developed AgriDEM software. Then a single pendulum cutting test device was built and dynamic cutting test was carried out to investigate the influence of straw-related physical properties and cutting parameters on the straw-cutting force. At the same time, a dynamic straw-cutting simulation was carried out based on the AgriDEM software. Comparing the simulation results with the actual test, the results showed that the simulation error of straw-cutting forces was increased with the increase of cutting speed when using quasi-static bonding parameters for dynamic cutting simulation. To reduce the simulation error at different cutting speeds, two calibration methods for the bonding parameters of wheat straws were proposed. The first method was a multi-variable central combination optimization calibration method with stiffness coefficient and critical stress as test factors, which obtained the optimized values of the normal stiffness coefficient, tangential stiffness coefficient, critical normal stress, and critical tangential stress for a cutting speed of 1m/s, respectively, as 2.82×1011N/m3, 1.32×1011N/m3, 9.50×106Pa and 6.46×106Pa, respectively. The second method was a single-variable calibration method with the bonding radius scaling factor, and the optimized values of the bonding radius scaling factor were obtained as 0.997, 1.106, 1.163, 1.213, 1.323 and 1.439 for the pendulum-cutting speeds of 0.5m/s, 1m/s, 1.5m/s, 2m/s, 2.5m/s, 3m/s, and a fitting equation between the cutting speed and the calibration value of the bonding radius scaling factor was obtained. Finally, by comparing the reciprocating cutting test with simulation at cutting speed of 1m/s, it was found that using the first calibration method with stiffness coefficient and critical stress, the simulation error of the straw cut-off force was reduced from 22.55% to 9.13%. Using the second calibration method, when the bonding radius scaling factor were 1.068, the simulation error of the straw cut-off force was reduced from 22.55% to 6.58%. This preliminary demonstrated that the two bonding parameter calibration methods proposed can effectively reduce the simulation error of straw-cut-off force, and the use of calibration value of the bonding radius scaling factor corresponded to the middle value of the effective cutting speed to simulate the reciprocating cutting operation was the most accurate and feasible method. The research result can provide a reference for the simulation of wheat cutting operation process and the optimized design of key cutting components.

Abstract:Aiming to address the current lack of fundamental models for studying the composite body system formed by intertwined multi-plant roots of rice pot seedlings within substrates, and establish a theoretical foundation for subsequent exploration of machine-pot body interactions during transplanting operations, a discrete element modeling method for rice pot seedling composite-body systems was proposed. The method integrated Matlab based root coefficient values to simulate growth patterns. Destructive testing was conducted to measure and analyze geometric morphologies, root topological relationships, and substrate-related parameters of rice seedling roots at the transplanting stage. By incorporating boundary constraints from seedling trays and root-root/root-tray interactive growth characteristics, characteristic functions describing root growth patterns were established. Matlab programming was employed to generate topological trajectories of root system growth. A segmentation sorting algorithm was developed to determine central coordinates of discrete particles. The EDEM software platform was utilized to integrate the established root-substrate discrete element geometric model with the EdinBurgh Elasto-Plastic Adhesion (EEPA) with Bonding contact mechanics model, thereby constructing the composite-body discrete element model. Validation through comparative compression and shear tests on rice pot seedling composite bodies demonstrated that simulation results exhibited consistent trends with experimental data, with errors meeting prescribed tolerances. This confirmed the feasibility of the proposed discrete element model for rice pot seedling composite-body systems.

Abstract:Aiming to address the lack of simulation models for cutting and crushing Neyraudia reynaudiana stalk, a discrete element model for Neyraudia reynaudiana stalk was established and the determination and calibration of specific parameters were completed. Firstly, a test platform was set up to determine contact parameters such as collision recovery coefficient, static friction, and rolling friction coefficient between the Neyraudia reynaudiana stalk and the steel plate. A Neyraudia reynaudiana stalk model was established, and a simulation stacking angle test was conducted. The contact parameters between stalks were calibrated by using the Box-Behnken test scheme and physical stacking angle (23.47°). Secondly, physical shear and compression tests were conducted to obtain the maximum shear force of the stem: Fd=155.18N, and the maximum compressive failure force: Fc=844.36N. To further establish the flexible Neyraudia reynaudiana stalk model, the Hertz-Mindlin with Bonding was selected as the inter-particle bonding model parameter. Simulation shear and compression tests of the model were carried out by using the Central Composite test program, and the bonding model parameter was calibrated with Fd and Fc as optimization targets. The parameters of the discrete element model for Neyraudia reynaudiana stalk were obtained as follows: collision recovery coefficients, static rubbing coefficients, and dynamic friction coefficients of the stalks were 0.35, 0.60, and 0.0049;the normal and tangential stiffnesses in the bonding model of the flexible Neyraudia reynaudiana stalk model were 6.13×1010N/m3 and 3.96×1010N/m3, respectively, and the normal and tangential stresses were 3.5×107Pa and 4.0×107Pa, respectively. The simulation validation test showed a maximum deviation of 4.81% between the physical and simulation test results, indicating that the discrete Neyraudia reynaudiana stalk model was more reasonable. Overall, the research result can provide a theoretical basis for the design of subsequent shear crushing devices for Neyraudia reynaudiana stalk.

Abstract:Aiming at the lack of internal bonding parameters of garlic species when using the discrete element method for the simulation and analysis of garlic precision sowing and harvesting and other key operational processes, Pizhou white garlic was selected as the research object, and a discrete element flexible model of crushable garlic species was established by using the EDEM software, and the bonding parameters of the species were calibrated. Taking the critical crushing load (105.5 N) and displacement (4.7mm) of the processed garlic species pieces without skin as the reference basis, the significance screening and optimization searching of the discrete element model of garlic species were carried out by Plackett-Burman, the steepest climb test and Box-Behnken test. The results showed that the optimal parameter combinations of normal stiffness per unit area, tangential stiffness, normal strength, shear strength and adhesion radius ratio were 5.115×108N/m3, 2.3×107N/m3, 1.265×106Pa, 54454.594Pa, and 1.716, respectively. At this time, the critical crushing load and displacement of simulated non-peeled garlic species pieces were 116.6N and 5.13mm, with the previous test without skinned garlic species, the error was 9.5% and 12.7%, and with the test with skinned garlic species (critical crushing load and displacement of 107.5N and 4.8mm, respectively), the error was 7.8% and 7.4%, which verified the accuracy of the model. The establishment of the discrete element model of garlic species can provide a reference for the subsequent vibratory species filling force analysis and attitude adjustment.

Abstract:To accurately represent the property differences in the vertical direction of sandy soil in northern open-field vineyards and address the lack of reliable discrete element simulation parameters in the study of interaction mechanisms with soil-touching components, simulation models for the upper, middle, and lower layers of sandy soil in open-field vineyards were established. The hysteretic spring contact model (HSCM) and linear cohesion model (LCM) were used as the contact models between soil particles. Firstly, the pile angle of each soil layer was obtained through cylindrical uplift tests combined with image processing technology. The rolling friction coefficients and restitution coefficients were measured by using rolling ball and collision rebound tests. With soil particle contact parameters as experimental factors and stacking angle as the indicator, a three-factor three-level experimental design was conducted to establish a regression prediction model for the soil stacking angle. The rolling friction coefficients for the three layers of soil were found to be 0.07, 0.17, and 0.21, respectively, with restitution coefficients of 0.47, 0.52, and 0.62, and cohesion energy densities of 2694J/m3, 4266J/m3, and 4432J/m3. Yield strengths of each soil layer were measured through soil yield tests. Secondly, the contact parameters between the soil and soil-touching components were calibrated. The sliding friction angles of each soil layer were determined through slope tests. With soil particle and soil-touching component contact parameters as experimental factors and sliding friction angle as the indicator, a general rotational center combination simulation test was conducted. Under the calibrated parameters, the errors between the simulated and actual stacking angles for the three soil layers were 1.7%, 2.6%, and 5.0%, respectively. To validate the accuracy of the calibrated parameters, field soil throwing tests were conducted, showing that the throwing distances for covered and uncovered sides of the soil were 101.4mm and 235.3mm, respectively, with errors of 3.34% and 8.73% compared with the simulation values. The relatively small errors indicated that the simulation model was accurate and reliable. The results can provide a theoretical basis for subsequent interaction analysis between soil and soil-touching components.

Abstract:Due to the lack of film model construction methods, the motion characteristics numerical simulation of the flexible materials on mechanical equipment is limited. A flexible model construction method based on a multi-feature comprehensive analysis of deformation, friction and aerodynamic characteristics was proposed, and its reliability was verified by using the plastic film as a study case. Firstly, the drape coefficient, static friction coefficient, friction angle, dynamic friction coefficient and suspension velocity of the film were measured to be 0.397, 0.417, 22.636°, 0.373 and 2.327m/s, respectively, by physical tests. Based on this, a film model was constructed by using the custom shell model of Rocky software. Taking the drape coefficient, friction angle and suspension velocity as the response indexes, the optimal parameter combination of the model was determined as the basic shell element area of 174.017mm2, elastic ratio of 0.850, density of 238.583kg/m3, Young’s modulus of 5.676MPa, film-film static friction coefficient of 0.437, and film-film dynamic friction coefficient of 0.397. Finally, the film group flow test was conducted to verify the general applicability of the model. It showed that the relative error between the simulated suspension velocity and the physical value was 3.18%, and the behaviour of migration, suspension, expansion loosening and disintegration were consistent with the physical test, indicating that the flexible model has feasibility and applicability.

Abstract:Pneumatic conveying is one of the main methods in the field of aquatic feeding. However, the mechanism of particle movement in the pneumatic conveying process is not fully understood at present, making it challenging to improve the operational efficiency of the pneumatic conveying system. The CFD-DEM gas-solid coupling numerical analysis method was utilized to construct a model for the pneumatic conveying process of granular feed, specifically focusing on systems with double bends. Additionally, the Box-Behnken response surface method was employed to quantitatively analyze the effects of inlet wind speed, feeding rate, and the number of bends on various parameters, including the ratio of granular feed material conveyed, the exit speed of particles, and the degree of levitation. The results of the sensitivity analysis indicated that the factors influencing the model response indexes in descending order of importance were as follows: inlet wind speed, feeding rate, and the number of bends. Specifically, the inlet wind speed was directly proportional to the velocity of the particle outlet and influenced the vertical coordinates and standard deviation of the particles. Notably, the inlet wind speed did not significantly affect the material-air transport ratio when the feeding rate was below 35g/s. As the inlet wind speed increased, both the exit velocity of particles and the degree of particle suspension were risen, suggesting an enhancement in the conveying performance of the pneumatic conveying system. Under optimal conditions, specifically, an inlet wind speed of 20m/s, a feeding rate of 27.232g/s, and a system without elbows-the material-air conveying ratio reached 0.966, the exit velocity of particles was 12.48m/s, the vertical coordinate of particles was -3.944mm, and the standard deviation was 8.805mm. The findings can provide a theoretical foundation for improving the efficiency of pneumatic conveying systems and optimizing their design.

Abstract:The rope model constructed based on the finite element method (FEM) has a complex structure and is mainly used for the mechanical analysis of rope structure, and its accuracy is greatly affected by the rope’s structure, resulting in poor model adaptability. The construction of structurally simple and adaptable rope models based on the discrete element method (DEM) is still a challenge. A rope modeling and parameter calibration method based on the DEM was proposed. The constructed rope model was cylindrical and contained triangular shell elements with different sizes and their joints. The forces and moments between the triangular shell elements were connected by a linear elastic model and a bilinear model, respectively. The DEM parameters of the rope model were calibrated by tensile and three-point bending tests with tensile strength and elastic modulus as response values, respectively. The optimal combination of the DEM parameters of the rope model was determined by solving the tensile strength and elastic modulus regression equations. The results of the validation tests showed that the tensile strength error and elastic modulus error of the rope model were 2.32% and 1.74%, respectively. The average errors of tensile and bending curves were 6.65% and 7.35%, respectively. Mechanical analyses showed that the DEM model can be used to analyze the local mechanical response of the rope during force application. The DEM-based rope model constructed was a single-strand structure. Compared with the FEM-based rope model, the structure was simple and universal.

Abstract:Aiming at addressing the deficiencies in the arch-back orientation function of existing strawberry transplanting devices, an automatic arch-back oriented transplanting apparatus for strawberry plug seedlings was designed. The system consisted of a seedling feeding device, a seedling extraction and orientation mechanism, a conveying device, and a vertical planting unit, which sequentially accomplished tray transportation, seedling extraction, arch-back orientation, and planting operations. Through analysis of morphological parameters of strawberry seedlings at transplanting stage, a pneumatic wedge-shaped elastic self-resetting extraction claw was developed to accommodate the characteristics of large strawberry plug seedlings. Motion parameter matching analysis for the vertical planting unit determined the optimal opening angle of the planting duckbill. Leveraging the inverse relationship between stolons and arch-back orientation, an orientation recognition method integrating ExG algorithm and DBSCAN clustering segmentation was implemented. Orthogonal experiments with insertion speed, extraction speed, and substrate moisture content as influencing factors revealed that substrate moisture (50%~60%) exerted the most significant impact on extraction success rate, followed by insertion speed (150mm/s) and extraction speed (124mm/s). System validation tests demonstrated that under optimal parameters, the device achieved a transplanting efficiency of 16 plants/min with 91.4% extraction success rate, 95.7% orientation accuracy, and 87.5% overall transplanting success. The average planting depth and spacing reached 98.8mm and 194.3mm, respectively, meeting agronomic requirements for strawberry cultivation. The research result can provide technical references for developing automated transplanting equipment in elevated strawberry cultivation systems.

Abstract:Hydroponic lettuce is the main crop in plant factories. However, mechanical harvesting currently causes serious leaf damage. The key problem of low-damage harvesting in plant factories was solved. A visual detection method for lettuce leaf expansion and overlapping points to control flexible grasping parameters was proposed, thereby improving harvesting quality. A vision-based leaf expansion detection method was carefully tested that used edge contour extraction to accurately measure leaf expansion, even when adjacent lettuce plants overlapped. The results were used to adjust flexible gripper diameters precisely. To optimize overlapping point recognition, comprehensive experiments were conducted by comparing YOLO v5s, YOLO v5s-SimAM, YOLO v5s-SENET, and YOLO v5s-CA models, with detailed analysis of their respective impacts on recognition accuracy parameters. Subsequently, based on the leaf expansion detection results, optimal overlapping points were screened. The positional information of these selected overlapping points was then used to calculate the most appropriate grasping angles for the robotic manipulator. Experimental results demonstrated that the visual detection system achieved an average relative error of merely 2.46% for leaf expansion measurement. Moreover, the YOLO v5s-CA model delivered superior performance in overlapping point recognition with 94.1% accuracy, 91.0% recall, and 93.8% mAP. Subsequent harvesting validation tests confirmed the effectiveness of this method, attaining a remarkable 97.23% success rate while maintaining minimal damage at just 4.08%, ultimately realizing high-quality, low-damage flexible harvesting of hydroponic lettuce.

Abstract:Tomatoes are one of the major crops in facility agriculture. Against the backdrop of a shortage of agricultural labor, the automated harvesting of tomatoes is of great significance. To meet the harvesting requirements of individual tomatoes in facilities, a smart harvesting platform for facility tomatoes was designed. The platform mainly consisted of a lifting mechanism, a harvesting mechanism, and an identification and positioning system. The overall harvesting structure was composed of a low-voltage integrated servo ball screw pair lifting mechanism, a six-axis collaborative robotic arm, and a force-controlled end effector, achieving automated, precise, and efficient harvesting operations. Based on an improved YOLO v5-HSV fusion algorithm for identification and detection, image threshold segmentation was performed on the H component. This improved the accuracy of identifying ripe target fruits and effectively eliminated the interference of unripe tomatoes and leafy backgrounds. Using the eye-in-hand calibration method, the ZED stereo camera was employed for positioning to obtain the spatial coordinates of target fruits in the robotic arm’s base coordinate system in real-time. In field harvesting experiments, the facility tomato harvesting robot prototype built achieved a recognition accuracy of 95.01%, increased the harvesting success rate to 87.96%, and reduced the average harvesting time per fruit to 14.56s. The results showed that the YOLO v5-HSV fusion algorithm can reduce recognition errors of tomatoes. The eye-in-hand algorithm was used to calculate the transformation matrix for accurate identification and positioning of target fruits. The recognition, positioning, and harvesting capabilities of the smart harvesting device for facility tomatoes met the practical operational requirements.

Abstract:In the process of seedling raising in the greenhouse cavity tray, there are phenomena such as seedling leakage and poor growth. In order to ensure the grade of factory seedlings, it is necessary to remove these inferior seedling substrate blocks and replant them with healthy seedlings. The existing end effector for seedling extraction is mainly suitable for large-seedling-aged bowl-body seedlings, and has poor adaptability to small-seedling-aged and weak-rooting bowl-body seedlings. Therefore, a flexible end effector was designed, which suitable for small space seedling planting and its row of multi-end seedling replenishment operation mechanism. The flexible end-effector can be controlled separately, and the claw blades were gradually retracted during the seedling picking process during the descending process, forming a progressive clamping operation on the substrate block. A bunch of red, caryophyllus, begonia, and petunia burrow seedlings were taken as the research objects. Carry out the rigid-flexible coupling dynamic analysis of the flexible end effector for manipulators. The orthogonal experiment of three-factor and three-level seedling transplantation of flexible planting end effector was carried out. The results showed that the experimental factors affecting the seedling root ball intergrity were claw spacing, claw width and substrate moisture content, and the optimal parameter combination was 21mm for claw spacing, 18mm for claw width and 55%~60% for substrate moisture content. Under the optimal combination of parameters, the end-effector seedling transplantation performance verification test was carried out. The results showed that the integrity of the bowl body was more than 85% after successful transplantation, which met the needs of seedling transplantation, and the average seedling replenishment efficiency of the whole machine was 2105 plants/hour (about 68 trays/hour). The research result can provide a reference for the automatic seedling replenishment operation of greenhouse cavity tray seedlings.

Abstract:At present, vegetable seedling grafting labor problems highlighted, and efficient grafting model which is suitable for our production was more scarce, thus a tomato vegetable vertical semi-automatic grafting machine was innovatively designed. The machine consisted of a six-station rotary table, synchronized cutting, seedling clamping, back to planting and other key institutions, scion and rootstock seedling position by using the upper and lower stacked vertical layout, a single person can be on the anvil scion seedling, the machine can automatically complete the seedling clamping cutting, clamping and back to planting into the hole operations. The rigid-flexible coupling model of the grafted seedling flexible clamping hand was established by ADAMS-ABAQUS, and the simulation analysis showed that the designed flexible clamping mechanism can effectively reduce the stress on the clamped seedling compared with the rigid clamping, and the maximum clamping force was reduced by 84.5%, which helped to realize low-loss clamping. Further three-factor and three-level orthogonal optimization test of the clamping end was carried out, and the optimal structural combination of parameters affecting the grafting success rate of the clamping end was found to be as follows: the thickness of the clamping shrapnel was 0.6mm, the working air pressure was 0.3MPa, and the length of the shrapnel connection was 10mm. At this time, the average success rate of the grafting machine was 97.3%, and the average grafting efficiency was 717 plants/h, which was twice as much as that of the manual grafting efficiency, and it can meet the requirements for the automated planting of eggplant and fruit vegetables in the factory. It met the demand for automated grafting in factory planting of eggplant and fruit vegetables.

Abstract:Automatic clamping is a critical step in mechanical grafting processes. To address the issues of unclear mechanical parameters of grafting clips for cucurbits and the potential impact damage caused to tender seedlings by instantaneous release, the complete clamping process was simulated by using transient structural finite element analysis. The dynamic mechanical characteristics of the grafting clip were analyzed to clarify how structural parameters affected the instantaneous impact force and stable clamping force on seedling stems during the clamping operation. A mathematical model with five factors was developed, using the clamp mouth displacement and opening angle at full closure of the grafting clip arms under varying steel ring diameters as structural parameters, and instantaneous and stable clamping forces as evaluation criteria. This enabled the theoretical calculation of clamping force for seedlings of arbitrary diameters. Utilizing a self-developed grafting machine with a progressive clamping mechanism, comparative experiments were conducted between gradual-release clamping at specified angles and traditional instantaneous release at maximum opening. Results showed that during the gradual-release clamping process at a specified angle, when the steel ring diameter was 0.7 mm, the instantaneous impact force on the seedling surface was 5.29N, below the safe compression limit for scions. This resulted in a 94% successful clamping rate without significant mechanical damage to seedling surfaces, thus satisfying safety and stability requirements for mechanical grafting clamping. The research result can provide theoretical support and references for the design and optimization of automated clamping mechanisms in grafting robots.

Abstract:In view of the problem that the seeds are easy to be squeezed and displaced after the transportation and placement of soft trays in vegetable seedling raising, a kind of dislocation stacking storage and transportation machine for seedling trays was designed. In order to avoid the compression of the matrix soil and the displacement of the seeds, a storage form of dislocation stacking of the tray was innovatively proposed, and the key components of the machine were designed and calculated, and the hardware selection and software design of the control system were completed. The selection range of each factor in the optimization test was determined by single factor test, and the Box-Behnken Design response surface optimization test design was carried out. The influence of four test factors, such as the dumping angle of the tray support plate, the height of the front end of the tray support plate, the descending speed of the screw slider in the translation-rotation joint conversion placement device and the moving speed of the whole machine, on the three evaluation indexes of the qualified rate of the declination pass rate, the longitudinal deflection distance and the transverse deflection distance was explored. The Design-Expert 13.0 software was used to optimize the parameters of the test results to verify the reliability of the optimized combination parameters. Considering the actual working parameters of the machine, the optimized pendulum parameters were set as the dumping angle of the tray support plate of 28.0°, the height of the front end of the tray support plate from the ground of 44.5mm, the descending speed of the screw slider of 38.5mm/s, and the moving speed of the whole machine of 59.0mm/s. The pass rate of above parameters were 93.3%, 93.3% and 94.4%, respectively. The absolute value of the error between the predicted value and the predicted value was less than 5%. The test and optimization results were basically the same, which met the agronomic requirements of the tray placement.

Abstract:Accurately predicting environmental parameters in solar greenhouses is crucial for achieving precise environmental control. In solar greenhouses, temperature, humidity, and light intensity are crucial environmental parameters. The monitoring platform collected data on the internal environment of the solar greenhouse for one year, including temperature, humidity, and light intensity. Additionally, meteorological data, comprising outdoor temperature, outdoor humidity, and outdoor light intensity, was gathered during the same time frame. The characteristics and interrelationships among these parameters were investigated by a thorough analysis. The analysis revealed that environmental parameters in solar greenhouses displayed characteristics such as temporal variability, non-linearity, and periodicity. These parameters exhibited complex coupling relationships. Notably, these characteristics and coupling relationships exhibited pronounced seasonal variations. The multi-parameter multi-step prediction model for solar greenhouse (MPMS-SGH) was introduced, aiming to accurately predict three key greenhouse environmental parameters, and the model had certain seasonal adaptability. MPMS-SGH was structured with multiple layers, including an input layer, a preprocessing layer, a feature extraction layer, and a prediction layer. The input layer was used to generate the original sequence matrix, which included indoor temperature, indoor humidity, indoor light intensity, as well as outdoor temperature and outdoor light intensity. Then the preprocessing layer normalized, decomposed, and positionally encoded the original sequence matrix. In the feature extraction layer, the time attention mechanism and frequency attention mechanism were used to extract features from the trend component and the seasonal component, respectively. Finally, the prediction layer used a multi-layer perceptron to perform multi-step prediction of indoor environmental parameters (i.e. temperature, humidity, and light intensity). The parameter selection experiment evaluated the predictive performance of MPMS-SGH on input and output sequences of different lengths. The results indicated that with a constant output sequence length, the prediction accuracy of MPMS-SGH was firstly increased and then decreased with the increase of input sequence length. Specifically, when the input sequence length was 100, MPMS-SGH had the highest prediction accuracy, with RMSE of 0.22℃, 0.28%, and 250lx for temperature, humidity, and light intensity, respectively. When the length of the input sequence remained constant, as the length of the output sequence increased, the accuracy of the model in predicting the three environmental parameters was continuously decreased. When the length of the output sequence exceeded 45, the prediction accuracy of MPMS-SGH was significantly decreased. In order to achieve the best balance between model size and performance, the input sequence length of MPMS-SGH was set to be 100, while the output sequence length was set to be 35. To assess MPMS-SGH’s performance, comparative experiments with four prediction models were conducted: SVR, STL-SVR, LSTM, and STL-LSTM. The results demonstrated that MPMS-SGH surpassed all other models, achieving RMSE of 0.15℃ for temperature, 0.38% for humidity, and 260lx for light intensity. Additionally, sequence decomposition can contribute to enhancing MPMS-SGH’s prediction performance. To further evaluate MPMS-SGH’s capabilities, its prediction accuracy was tested across different seasons for greenhouse environmental parameters. MPMS-SGH had the highest accuracy in predicting indoor temperature and the lowest accuracy in predicting humidity. And the accuracy of MPMS-SGH in predicting environmental parameters of the solar greenhouse fluctuated with seasons. MPMS-SGH had the highest accuracy in predicting the temperature inside the greenhouse on sunny days in spring (R2=0.91), the highest accuracy in predicting the humidity inside the greenhouse on sunny days in winter (R2=0.83), and the highest accuracy in predicting the light intensity inside the greenhouse on cloudy days in autumm (R2=0.89). MPMS-SGH had the lowest accuracy in predicting three environmental parameters in a sunny summer greenhouse.

Abstract:Accurate prediction of environmental parameters in solar greenhouse can effectively reduce the lag of greenhouse control, which is the key to realize the intelligent control of solar greenhouse. It is of great significance to realize the fine intelligent management of solar greenhouse production process. A multi-layer LSTM network model was set up and combined with SVM to design a hybrid model for predicting greenhouse environmental parameters. In the process of model prediction, weather information in the future period was added to supplement the modified model to improve the accuracy of model prediction results and realize the prediction of environmental parameters in greenhouse under complex weather conditions. The experimental results showed that the model made full use of the advantages of LSTM in long-term dependency processing of sequence data and the advantages of SVM in feature extraction and classification. The accurate prediction of the key environmental parameters in the solar greenhouse under the complex outdoor environmental conditions was realized, and the prediction determination coefficient R2 was not less than 0.93. The experimental verification and comparison showed that the proposed LSTM-SVM hybrid prediction model with future weather information correction was superior to the single model and other traditional methods in the prediction accuracy and stability of solar greenhouse environmental parameters under complex weather conditions.

Abstract:Dynamic regulation of nutrient solution is an important part of soilless cultivation systems. Monitoring the nutritional status of reused solutions to ensure effective supplementation of nutrients to the optimal concentration is crucial for achieving recycling and reducing economic costs. However, online detection of nutrient solution components is challenging. A soft sensing method for other major ions was proposed based on pH, EC, nitrate and ammonium ion activities, and corresponding control software and hardware systems were designed. Firstly, the entire cycle of hydroponic lettuce was conducted by using a universal formula for horticultural experiment. Based on the calculation formulas for ion activity and EC contribution rate, the variation patterns of nutrient solution components were studied. The results showed that the effect of ion activity on nutrient solution EC was stackable and the contribution rate of ion EC was relatively stable. Support vector regression (SVR) and genetic algorithms (GA) were used to establish Ca2+, Mg2+, and K+ion activity regression models. The optimal determination coefficient and root mean square error of the models were 0.9651, 0.9724, 0.9540, and 0.0146mmol/L, 0.0262mmol/L, 0.1437mmol/L. Secondly, a multi ion concentration control strategy for the universal formula of garden trial leafy vegetables was designed. Finally, a control system for hardware and software was established to achieve detection and supplementation of water and fertilizer status. The validation experiment showed that the nutrient absorption was relatively small in the 0~10 days after transplanting, and the control instruction was executed on the 5th day after adding liquid;the nutrient absorption rate was increased rapidly in the 10~30 days after transplanting, and the control was carried out based on the increase of potassium and nitrogen elements that were obviously absorbed;in the harvest period, the total nutrient absorption was increased, but the absorption rate was decreased significantly, and the control was carried out based on the calcium element that had the fastest relative absorption rate. The results showed that the research improved the EC control level of nutrient solution to the level of ion concentration control, providing theoretical basis for meeting the dynamic nutrient solution ion concentration requirements of crop production.

Abstract:As an advanced stage and manifestation of facility agriculture, plant factories facilitate highly controllable and uninterrupted industrial production of crops throughout the year. With advancements and applications in plant synthetic biology, systematic design and modification of plant genes, metabolic pathways, and physiological functions through engineering approaches are now being integrated with plant factories in various areas, including the production of plant secondary metabolites, molecular agricultural products, enhancing resource utilization efficiency, and the creation of carbon sink plants. This integration is transitioning from the laboratory to industrialization. The high-value sustainable development of plant factories and the collaborative innovation between industry and technology was delved. It elaborated on the combined applications of plant synthetic biology and plant factory production technology in fields such as bio-manufacturing, drug development, endangered plant conservation, and future space stations. Furthermore, it explored the future directions and challenges of this integration, offering novel solutions to expand the boundaries of agricultural production possibilities, achieve efficient supply of high-value agricultural products, and foster diversified development of agricultural resources.

Abstract:Aiming at the instability issue of seed transportation under high-speed seeding operations, a wheel-type seed guide device was proposed. Using a high-speed precision seed-metering device for maize as the test platform, the operational principle of this guide device was elucidated. Furthermore, the mechanism affecting seeding performance was systematically analyzed through kinematic modeling and discrete element simulations. In a single-factor test, the effects of the circle centre angle of the constrained transport arc surface at the base of the wheel-type seed guide device and the rotational speed of the seed guide wheel on the seed metering performance were investigated. High speed camera technology was used to analyze the trajectories and discharge velocity of maize seeds at various levels of factors. The findings demonstrated that an elevation in the circle centre angle of the constraint transport arc surface enhanced the stability of maize seed discharge, resulting in a more uniform discharge velocity. A multi-factor test was used to determine the optimum operating parameters of the wheel-type seed guide device. The optimal parameters were determined to be a circle centre angle of the constraint transport arc surface of 65° and the seed guide wheel rotational speed of 120r/min. Multi-factor test results indicated that the qualified rate was 98.35%, the coefficient of variation was 8.13%, the multiple rate was 1.18%, and the miss rate was 0.47% under the optimal operating condition. This research can provide a reference for the investigation and development of the mechanical seed guide device, and also a reference for investigating the mechanism for improving high-speed seed metering performance.

Abstract:During cotton field sowing, missed seeding frequently occurs, and manual monitoring struggles to ensure accuracy while imposing high labor demands. However, current research on seeding monitoring for cotton precision hill-drop planters remains limited, particularly in effectively tracking the seed pickup status of hill-drop planter. To address these challenges, a multi-source data monitoring method was proposed based on coupled fiber optic and inductive sensing, which offered superior accuracy and installation flexibility compared with traditional approaches. The developed monitoring system employed fiber optic sensors and inductive sensors as signal acquisition sources, with LabVIEW software as the control platform and an industrial computer as the monitoring terminal. The hardware and software designs of the system were completed, including signal filtering and data acquisition/calculation modules, enabling real-time monitoring of critical sowing parameters such as seeding quantity, missed seeding count, cavitation number, missed seeding rate, and operational speed. Bench tests calibrated the light intensity levels and thresholds of the fiber optic sensors, identifying optimal settings under black seed pickup disc conditions. The inductive sensors achieved a monitoring accuracy exceeding 98.21% for hill-drop counts in bench tests. Field trials demonstrated that the system maintained a hill-drop count accuracy no less than 98.10%, seeding quantity accuracy no less than 98.21%, and missed seeding accuracy no less than 95.00% at forward speeds of 2.5km/h to 3.8km/h. The high-precision monitoring system proved suitable for practical sowing operations, offering significant potential for enhancing cotton sowing intelligence and improving cotton industry efficiency.

Abstract:In view of the long transportation time of chemical fertilizer particles during the fertilizer discharge operation of the spiral fertilizer discharger, which led to the problem that the lag distance of fertilization in the initial stage of fertilization operation affected the accuracy of fertilization amount, an inclined spiral fertilizer discharger was designed, and the movement time of chemical fertilizer in the fertilizer discharger through the accelerate fertilizer fertilizer measure of the fertilizer discharger was reduced, and the lag distance of fertilization to improve the accuracy of fertilization was shortened. EDEM was used to simulate the movement of fertilizer particles in the fertilizer discharger under the conditions of accelerate fertilizer and unfilled fertilizer. Under the condition that the oblique angle of the fertilizer discharger and the maximum speed was 191.74r/min, the full speed of the fertilizer particles was increased from 0.017m/s to 0.162m/s, and the vertical velocity was increased from 0.005m/s to 0.057m/s, which increased the speed of the fertilizer particles leaving the fertilizer discharger and shortened the transportation time of fertilizer particles by 3.199s. The bench test was carried out. The test results showed that pre-filling fertilizer could significantly shorten the lag distance of fertilization, but the lag distance under pre-filling fertilizer was increased with the increase of operation speed. The lag distance of unprefilled fertilizer was increased with the increase of fertilizer application rate, but it had no obvious relationship with the operation speed. The amount of fertilizer applied under the condition of accelerate fertilizer was higher than that without accelerate fertilizer, which increased the accuracy of fertilization, but there was an overshoot of fertilizer application, which can be solved by further optimization. The field experiment was carried out by modifying the seeder. The field experiment results showed that the average lag distance of fertilization was 1.06m when the pre-filling fertilizer was not pre-filled, and the average lag distance of pre-filling fertilizer was 0.40m, which was shortened by 0.66m. The precision of fertilization control was improved from 77.6% to 96.5%.

Abstract:In response to the insufficient conveying capacity of intermediate conveying devices in current corn grain harvesters when paired with 13 rows or more of corn headers operating at feeding rate of 16kg/s, as well as the tendency for clogging during fluctuations in feeding volume, a variable gap conveying device was designed. The working principle of the device was outlined and parameter design and motion analysis for key components such as the floating mechanism and chain-type conveyor mechanism were conducted. A virtual prototype model of the variable gap conveying device was established by using the multi-body dynamics software RecurDyn. The physical models of corn ears were obtained through 3D scanning technology, and a discrete element model (DEM) of corn ears was subsequently created in the discrete element software EDEM. By utilizing EDEM-RecurDyn coupling simulation, the trajectory of corn ears within the variable gap conveying device, as well as their velocity and force changes during various conveying stages were analyzed, which clarified the conveying laws of corn ears and identified the primary factors influencing conveying performance. A three-factor, three-level simulation experiment was conducted, with spiral conveyor speed, conveyor sprocket speed, and conveying gap as the experimental factors, and corn ear conveying rate and damage rate as the evaluation indicators. The optimal operating parameters obtained from the simulation experiments were then used to conduct field trials. The field test results indicated that when the feeding rate was 16kg/s, spiral conveyor speed was 160r/min, conveyor sprocket speed was 460r/min, conveying gap was 60mm, threshing drum speed was 400r/min, concave clearance was 30mm, fan speed was 1200r/min, vibration frequency was 5Hz, and upper sieve opening was 18mm, the variable gap conveying device achieved a conveying rate of 98.98%, with grain breakage rate of 1.37%, impurity content of 0.85%, and total loss rate of 1.09%. The field validation tests were in good agreement with the simulation results, fulfilling the operational requirements of the conveying device.

Abstract:In response to the problem that the potato picking device installed on a self-propelled potato picking and bagging machine is prone to soil accumulation in the front, resulting in poor transportation of the potato soil mixture. Therefore, the driver needs to frequently manually adjust the depth of the picking shovel into the soil, which can easily lead to misjudgment and result in the failure of the key operation indicators of the whole machine’s potato damage rate, leakage rate, and impurity rate to meet the operation requirements. A potato picking device operation parameter self adjustment system was studied. During the operation of the device, based on the real-time measurement of the height of the soil accumulation by ultrasonic sensors, the depth of the picking shovel entering the soil, the linear speed of the conveying chain, and the vibration intensity of the potato soil separation were continuously adjusted through hydraulic drive to achieve optimal operation quality. Based on DEM-MBD coupling simulation, the simulation was conducted on the operational quality of the picking device under different combinations of operational parameters, clarifying the general law of the impact of operational parameters on operational quality. Based on this law, a self adjustment method for homework parameters was designed. Based on the control requirements of device operation parameters, a simulation of electro-hydraulic control system was conducted by using AMESim-Matlab/Simulink. The control effects of fuzzy PID and classical PID were compared, and the simulation results showed that fuzzy PID control improved the response speed of valve-controlled cylinder system and valve-controlled motor system compared to classical PID control. The field experiment was carried out by using the fuzzy PID control operation parameter self adjustment system on the picking device. The test results showed that the picking device equipped with the operation parameter self adjustment system had a potato leakage rate of 2.17%, a damaged potato rate of 2.51%, and a impurity rate of 1.64% during operation. All indicators met the operational requirements and were superior to manual control, verifying the superiority of the potato picking device equipped with a self adjusting system for operational parameters over manual control.

Abstract:In response to the problems of complex terrain in domestic tea gardens, irregular canopy shapes, and poor quality of machine-harvested bulk tea, a self-propelled contour-following tea harvester with adjustable cutter position was designed based on the growth characteristics of bulk tea canopies and the technical requirements for harvesting. The technical requirements for contour-following bulk tea harvesting were determined, and the overall structure and working principle of the self-propelled contour-following bulk tea harvester were analyzed.The contour-following harvesting device, pneumatic conveying device, transmission system, and contour-following control system were designed to make the whole machine more compact and stable in operation. To improve the quality of fresh tea leaves harvested in contour-following mode, the cutter was further optimized. Through computer simulation and response surface experiments, the cutting performance was evaluated based on the area ratio of different cutting regions in the cutting diagram. The optimal combination of cutter parameters was determined as a front bridge width of 3.44mm, a cutter tooth height of 20.63mm, and a cutter-to-machine speed ratio of 1.2. Under this parameter combination, the missed cutting rate was 3.90%, and the over-cutting rate was 16.36%. This effectively reduced the missed cutting rate while maintaining a relatively low over-cutting rate, indicating the effectiveness of the parameter optimization. Field trials verified that the contour-following harvester had good operational performance. During the trial period, the average intact leaf rate was 85.66%, the average missed collection rate was 1.17%, and the average missed cutting rate was 0.79%. All three indicators met the industry standards for tea harvester operation quality. Additionally, the canopy condition of the tea trees after operation was good, which did not affect subsequent tea production and met the design requirements of the tea harvester.

Abstract:Spinach, as a typical stem and leaf vegetable, has relatively high economic and nutritional value. To solve the problems such as poor orderliness and large damage to stems and leaves existing in the mechanical harvesting operation of spinach, a spinach ordered harvester with root cutting and gathering under the soil was designed. The collection scheme of cutting the roots of four rows of spinach under the soil and gathering them into two rows for orderly conveying was adopted, reducing the size of the lateral collection frame. It can achieve the operation process of “clamping into the soil and cutting roots-orderly conveying-gathering and collecting-turning and spreading” at one time. The root cutting shovel mechanism and the clamping, conveying, gathering and collection device were designed and analyzed. The key parameters of the triangular root cutting shovel and the clamping gap of the clamping belt that prevented spinach from dropping and stem damage during clamping and conveying were determined. A theoretical analysis was conducted on the changes in plant posture during the instantaneous working process of clamping and cutting, and the main factors influencing the posture changes of spinach plants were obtained. A bench test was conducted by using the Box-Behnken experimental design method. A three-factor and three-level orthogonal test was carried out with the machine’s forward speed, the inclination angle of the clamping track, and the conveying speed of the clamping belt as factors, and a quadratic regression model was established with the harvest loss rate and the spinach stem injury rate as response indicators. The influence of each factor on the response index was analyzed by the response surface method. The verification test results showed that when the machine’s forward speed was 1.1km/h, the inclination angle of the clamping track was 18°, and the conveying speed of the clamping belt was 0.45m/s, the harvest loss rate was 2.47%, the stem damage rate was 3.05%, and all performances met the requirements of mechanized spinach harvesting.

Abstract:With regard to the drawbacks of large structural dimensions and heavy weights in the current end-effectors for broccoli harvesting, a six-bar end effector for broccoli harvesting was designed, which was compact in structure, small in motion space and light in mass. Initially, the physical features of broccoli were measured, and the cutting force of broccoli stem tissues was evaluated and analyzed. Then, the incipient cutting angle of the cutter was set as 90°, and the cutting force of the cutter was specified as 100N. Afterwards, based on the physical parameters and growth situations of broccoli, the plan of the six-bar end-effector for harvesting was settled. The kinematic and dynamic-static mathematical models of the end six-bar mechanism were erected. By applying the genetic algorithm and taking the minimization of the external driving force and the minimization of the difference between the theoretical track and the actual track as the goals, the parameters of the end effector were optimized via a dual-objective optimization method. After optimization, the results of the bar lengths of the six-bar structure and the minimum driving torque of the end six-bar mechanism were achieved, with the torque being 1.37N·m. The optimized cutting track of the cutter was procured, and the optimization difference of the track was 1.29mm. The distance between the unilateral cutter and the largest flower cluster was 25mm. Finally, according to the optimized results, a servo motor with a rated power of 400W was picked as the power supply. A prototype was designed and produced following the optimized mechanism parameters, and experimental researches on the harvesting end effector were performed. The success rate of the harvesting trials of the end effector was 92.5%. The experimental outcomes demonstrated that the six-bar end effector for broccoli harvesting had a good harvesting performance.

Abstract:In order to ameliorate the current situation of high labor intensity and cost of manual harvesting, lack of mechanized harvesting technology and equipment in Zizania latifolia industry, the key technologies of mechanized harvesting of Zizania latifolia were developed based on the planting agronomy, plant characters and requirements of mechanized harvesting. A crawler self-propelled single-row harvester for Zizania latifolia was developed which can complete the cutting, conveying, blade cutting, and collection processes in one go, achieving mechanized harvesting of Zizania latifolia. The structure and working principle of the harvester were elaborated, and the key components such as tracked crawler power chassis, double-disc cutting device, clamping and conveying device, vertical conveying and blade cutting device were designed based on the theoretical analysis. In addition, the main structural and working parameters of the harvester were determined. Finally, the Zizania latifolia harvester was developed, and field experiments were carried out. Field test results showed that the driving performance and passing ability of the Zizania latifolia harvester in the Zizania latifolia field were good. When the working speed was approximately 0.4m/s, the clamping and conveying speed was approximately 0.7m/s, and the motor driving speed of the double-disc cutting device was 500r/min, respectively, the success rate of cutting was 100%, the average total loss rate was 5.2%, the average damage rate of shelled Zizania latifolia was 7.3%, and the average work efficiency was 0.14hm2/h. Moreover, the working parts of the harvester worked stably and showed a good effect, which indicated that the harvester prototype can basically meet he requirements for mechanized harvesting of Zizania latifolia. The research result can provide ways for the design and improvement of mechanized harvesting equipment for Zizania latifolia.

Abstract:Due to the high resistance of deep-root excavation, high energy consumption for large soil separation, and low efficiency in licorice harvesting, the variable amplitude oscillation licorice harvester was designed. Combined with the conditions of direct seeding licorice harvesting and the basis of previous research, the working conditions of the excitation device, excavation and separation device and rotary cutting device were clarified, and the dimensions of each key component were determined. The structure model of licorice harvester was constructed, and the working principles of the key components were analyzed. Response curves of trajectories and throw coefficients of the shovel fence working surface were solved, the kinetic equation of the shovel fence working surface was established, and the key factors affecting the licorice harvesting effect were identified. According to the soil breaking theory, the energy consumption equation of soil unit breaking through the shovel fence working surface was solved, and the mathematical model of additional breaking area and working parameters was obtained. The movement characteristics of the rotary cutting device and its cutting action on the surface licorice were analyzed. The results of the field test showed that the smooth working of the variable amplitude oscillation licorice harvester enabled vibratory crushing of the licorice-soil complex, transport and efficient separation of the licorice-soil mixture. At an amplitude of 10mm, a vibration frequency of 9Hz, and a forward speed of 0.07m/s, the licorice harvester excavation depth was 609mm, the licorice harvesting rate was 96.38%, the traction resistance was about 95kN, the effective drive torque was about 1185N·m, and the total working power was about 75kW. The research result can provide a method reference for the development of energy-saving and efficient licorice harvesting machine.

Abstract:At present, domestic cotton stalk crushing and residual film recovery machinery is characterized by lack of self-propelled models, poor combined operation performance, low production efficiency, poor film and stalk separation, low cotton stalk baling density and baling rate, low residual film collecting ratio, low residual film packaging density and baling rate, therefore a self-propelled combined operation machine for cotton stalk baling and residual film packing was designed. It was made up of the steering axle, front drive axle, hydraulic system, cab, frame assembly, power assembly, residual film collection and packing assembly, and cotton stalk crushing and baling assembly. The frame assembly was of upper and lower structure, i.e., including the upper frame and lower frame, and the cotton stalk crushing and baling assembly, cab, drive axle, power assembly, residual film recycling and baling assembly and steering axle were mounted on the frame from front to back in sequence. Cotton stalk crushing and baling assembly included cotton stalk crushing device, cotton stalk lifting device and cotton stalk baling device. Residual film collecting and packaging assembly included residual film picking device, residual film conveying device and residual film packing device. Hydraulic system included running system and working system, and the machine used fully hydraulic drive, which can collect and control the operating parameters of cotton stalk crushing, conveying and baling, and residual film picking, conveying and packaging for real-time, so as to avoid the occurrence of faults and improve the performance of the machine. Double screw Y-type blade cotton straw crushing device, scraper conveying device with front and rear dials and piston type cotton stalk baling device were used for cotton stalk crushing and collecting. Spiral discharge device, adjustable height loosening teeth, nail type film pickup device and coreless roll residual film packing device were used for residual film collecting and packing. It can achieve the separation of film and stalk, stalk crushing, stalk conveying and compressed baling, film pickup, film conveying, and packing at the same time. Field tests showed that when the operating speed was 7~10km/h, the depth of the loosening film tooth was 40~60mm, and the surface state of the packing belt was rough and the linear speed of the packing belt was 1.2m/s, the straw loss rate was 3%, the qualification rate of cotton stalk chopping was 97%, the average density of the stalk bale was 120kg/m3, the balling rate of stalk was 98.5%, the residual film collecting ratio was 90%, the average density of the residual film bale was 110kg/m3, the balling rate of the residual film was 99%, the separation ratio of cotton stalk and residual plastic film was 90%, the operation efficiency was 1.4hm2/h. The research result can provide effective equipment for the centralized treatment and reuse of stalk and residual film.

Abstract:Aiming at the problems of high labor intensity, low operation efficiency and uneven spraying of bud inhibitor in the process of cigar tobacco leaf planting and production, the work process of clamping, shearing, and spraying in turn was proposed. And the knapsack topping and bud suppression machine for cigar tobacco was developed. The machine was mainly composed of synchronous clamping mechanism, shearing mechanism and bud inhibitor spraying device. The mechanical analysis of the synchronous clamping mechanism under stable clamping state was carried out, and the key structural parameters were determined. Combined with the analysis of the structural characteristics of the top bud and the principle of sliding cutting, the key structural parameters of the shear blade were obtained. The spraying device of the bud inhibitor adopted the spraying operation mode of annular jet. According to the internal flow characteristics of the jet nozzle, the range of key structural parameters of the spraying device was analyzed. Based on Fluent software, the single factor simulation test of the nozzle flow characteristics was carried out. It was concluded that when the diameter of the main channel was 5~7mm, the number of nozzles was 5~7, and the diameter of the nozzles was 1~2mm, the fluid velocity between the nozzles was more uniform. The field experiment of spraying device structure optimization was carried out with the pesticide rate of cigar tobacco leaves as the index. The results showed that when the diameter of the main channel was 6mm, the number of nozzles was 6, the diameter of the nozzle was 1.5mm, and the spraying angle was 60°, the pesticide area of cigar leaves was 0.37%. The field performance test results showed that the average top bud cutting rate of the cigar tobacco leaf topping and bud suppression device was 92.40%, the top bud drop rate after cutting was 3.20%. And when the spraying flow rate of the bud suppressor was 25mL/s, the average axillary bud infection rate was 93.75%. The knapsack topping and bud suppression machine for cigar tobacco can complete the process of topping and bud inhibition, and the bud inhibitor was sprayed evenly. The research can provide an approach for the development of cigar tobacco leaves topping and bud inhibition machine.

Abstract:Aiming to address the issue of high breakage rates during the harvesting of high-moisture corn kernels, a threshable simulated electronic corn ear was designed. It was embedded with an inertial measurement unit (IMU) module and a flexible film pressure sensor, which can detect the dynamic impact force and static squeezing force during the threshing process in real time. Furthermore, it provided a technical method for exploring the principle of grain breakage during the threshing process. To verify the working performance of the simulated electronic corn ear, dynamic impact force test and static squeezing force test were conducted separately. The results of the dynamic impact force detection test showed that the average error of the impact force was 0.91N, the maximum error was 2.25N, and the average detection accuracy was 98.15%. The results of the static squeezing pressure detection test showed that the average detection error of the squeezing pressure was 2.47N, and the maximum average detection error was 7.25N. The curve fitting degree with the universal testing machine was high, and the R2 was 0.9874. Finally, threshing tests on the simulated electronic corn ear under actual working conditions were conducted by using a longitudinal flow threshing device. The experimental results indicated that the electronic corn ear could effectively detect its stress parameters in the threshing device. Under the conditions of a feeding rate of 2.5kg/s, a threshing drum speed of 450r/min, a threshing gap of 40mm, and a guide angle of 45°, the average impact force on the electronic corn ear was 17.40N, and the maximum impact force was 70.73N. The average squeezing force was 34.49N, and the maximum squeezing force was 96.30N. The research result can provide a technical means for analyzing the stress parameters of corn ears under high-speed flow group threshing conditions, and provide a new research method for exploring the principle of grain breakage.

Abstract:Quality detection of sugarcane seed segments is a key step in the pre-cutting preparation process. To address challenges such as complex environments and limited computational resources on edge devices in practical seed preparation scenarios, a lightweight sugarcane seed quality detection method was proposed based on an improved YOLO v8n model. By introducing mixed local channel attention (MLCA) and partial convolution (PConv) to the network structure, redundant computations were effectively reduced while enhancing the model’s focus on critical features. A dynamic upsampling module (DySample) was adopted in the neck network to more accurately capture the boundaries and details of feature points. Additionally, the detection head was improved by using a re-parameterized shared convolution structure, further reducing model complexity. The final model, named YOLO v8-SD, was integrated with a monocular camera distance estimation algorithm to calculate the distance between the cut surface and adjacent sugarcane nodes. Experimental results showed that YOLO v8-SD achieved an mAP50 of 98.3%, a frame detection speed of 142.9f/s, and a model size of only 3.45MB. Compared with the original model, the parameter count and FLOPs were reduced by 47.8% and 33.3%, respectively, the model size was reduced by 41.9%, and the frame rate was increased by 7.8f/s. Furthermore, the proposed method achieved an average relative error of less than 6.1% in cut-surface distance estimation. The improved model was deployed on an NVIDIA Jetson Orin NX development board and tested in a prototype sugarcane seed screening system. The results demonstrated an average detection accuracy of 97.33% across different sugarcane varieties, meeting the requirements of practical applications.

Abstract:The interaction between sediment particles and turbulent two-phase media inside hydro-turbine under sand containing water flow conditions has a significant impact on the erosion and wear mechanism of the turbine wall. Based on GPU accelerated coupled particle discrete element (DEM) and computational fluid dynamics (CFD) methods to solve the internal particle fluid system of hydro-turbine, the wear characteristics of sediment particle flow in hydro-turbine through DEM-CFD coupling simulation were obtained by using the real shape modeling of sediment particles. The results showed that GPU acceleration can save computational costs. Sediment particles under coarse sand flow were affected by the water flow and moved in a “short spiral” manner in the draft tube at small opening case, while at big opening case, they moved in a “long spiral” manner. The maximum resistance of spherical polyhedral sediment particles at small opening case was 34.78% higher than that of spherical particles. The collision strength of particles on the guide vanes at big opening case was 74.13% higher than that of the stay vanes, indicating that their impact wear was stronger than that of the stay vanes. The average maximum sliding distance of particles on the stay vanes was 21.43% higher than that of the guide vanes, indicating that their shear wear was stronger than that of the guide vanes. The wear of the guide vanes, runner blades and draft tube walls was consistent with the experimental results in the literature and the actual wear of the power plant. It was beneficial to control the operating case of the hydro-turbine during the flood season and operate it at small opening case. The research result can provide theoretical reference for improving the sediment erosion and damage of hydro-turbine.

Abstract:Poultry farming is an important sector of animal husbandry. The farming methods are moving towards digitalization and intelligence. Accurate and efficient perception of poultry information is the foundation of smart poultry farming. The development of artificial intelligence technology has provided new opportunities. The continuous improvement of information collection equipment has also played a key role. These advancements have enabled precise and efficient perception of multimodal health information in poultry. The research progress in poultry health monitoring was reviewed. The studies were based on multiple modalities, including images, acceleration data, wireless radio frequency, and audio signals. The focus was on three aspects: external morphology, behavior, and body temperature. A systematic analysis was conducted on poultry health monitoring devices and technologies. The advantages and disadvantages of different sensing modalities were summarized. The applicability of various collection devices was considered. Challenges in current research were discussed. Finally, future trends in intelligent poultry health monitoring were outlined. The key directions included improving algorithm accuracy, real-time performance, and robustness. The research aimed to provide methods and ideas for poultry health management and the sustainable development of the poultry industry.

Abstract:To enhance the quality of commercial fruits, it is imperative to detect mechanical damage at an early stage. Aiming to addresse the issue of unclear spectral evolution mechanisms in the early stages of fruit damage, leading to low accuracy and robustness of the damage detection model. The temperature difference time series data of damaged and undamaged areas in the early stages (0~2h) of mechanical damage was collected to Red delicious apples, and combined it with the spectral features of sensitive bands (1255~1314nm) by using the Granger causality test method to reveal and substantiate the mechanism of spectral feature evolution caused by temperature changes: with the progression of apple damage, the temperature difference was firstly risen, then fell, and finally stabilized. The more severe the damage was, the greater the temperature difference would be. The results of the Granger causality test indicated that the main cause of the change in the spectral features of the apple’s damage site in the 1255~1314nm characteristic band was the change in temperature at the site of damage. The milder the damage was, the weaker the influence of the temperature difference on the spectral features would be. The spectral features R were more representative and comprehensive in characterizing the surface temperature of the apple after damage. Through the study of the evolution mechanism of spectral features after the Red delicious apples was bruised, the research result can provide a theoretical basis and research approach for more reliable early detection of minor apple damage.

Abstract:The asynchronous maturity of tobacco leaves in natural environments presents a significant challenge for selective harvesting. Existing algorithms for tobacco leaf maturity classification often lack robustness under complex lighting conditions and struggle to balance model light weightness with classification accuracy. To address these issues, a deep learning-based approach that integrated a visual saliency algorithm with adaptive illumination correction to achieve accurate maturity classification of tobacco leaves in field environments was proposed. Firstly, the lightweight YOLO v8n model was used to detect tobacco leaf regions in the input image. Then, adaptive Gamma correction was applied to these regions to mitigate the impact of varying illumination on maturity classification. Next, an improved visual saliency algorithm was employed to generate saliency maps of the tobacco leaves. Finally, an enhanced MobileNetV2 network was used to classify the maturity levels of the detected leaves. Experimental results demonstrated that the lightweight YOLO v8n model reduced memory usage to 29% of the original model while achieving a mean average precision (mAP@0.5) of 94.8%. When using RGB images as input, the proposed method attained a classification accuracy of 91.93%. After integrating the saliency enhancement algorithm, the accuracy was increased to 95.07%, outperforming the current state-of-the-art model (MobileNetV3Large) by 2.03 percentage points. Additionally, the use of adaptive Gamma correction increased the information entropy of tobacco leaf images by 7.5% under high illumination conditions. These results validated the effectiveness of the proposed method, which combined adaptive Gamma correction with visual saliency enhancement. The approach offered a promising solution for improving tobacco leaf maturity classification and provided technical support for the development of tobacco harvesting robots.

Abstract:Estimation of chlorophyll status at jointing stage of winter wheat is important for nutritional diagnosis of winter wheat. The UAV remote sensing platform was used to obtain the remote sensing information of winter wheat growth at the jointing stage, the multispectral vegetation indexes, RGB image texture features and coverage information were extracted, the models for estimating the winter wheat SPAD values were constructed based on the multivariate linear regression (MLR) and random forest regression (RFR). Then the effects of multispectral vegetation indexes, textural features, cover information, and combining them with each other on the estimation of winter wheat SPAD values were analyzed. The results showed that the combination of multispectral vegetation indexes, texture features, canopy coverage (combination of two or three types of parameters) can be used for the estimation of winter wheat SPAD values at the jointing stage, and the combination of more types of parameters improved the estimation accuracy of winter wheat SPAD values compared with the combination of single-type parameters or two types of parameters. And the accuracy of the winter wheat SPAD estimation model constructed based on the same parameters by using RFR was higher than that of the model constructed by MLR. Among them, the model constructed based on the three types of parameters had the highest estimation accuracy of winter wheat SPAD values, with R2 of 0.78 and RMSE of 2.08. Moreover, the effects of each type of parameters on the models accuracy improvement in descending order were multispectral vegetation indexes, texture features, and canopy coverage. Among them, the accuracy of the model constructed by multispectral vegetation indexes was similar to that of the model constructed by texture features. Canopy coverage had the smallest improvement in the estimation accuracy of SPAD values, but combining other features could improve the estimation accuracy of winter wheat SPAD values (R2 was increased by 0.02~0.03 for the RFR models). The combination of multispectral vegetation indexes, texture features and canopy coverage improved the accuracy of the models, providing a fast technical reference solution for winter wheat SPAD values estimation.

Abstract:In the process of using rice seeds, it was essential to classify them, but this problem was highly challenging because rice seeds have similar visual appearance. To address the challenges associated with recognizing multiple rice seed varieties, specifically the abundance of morphological features and the high classification complexity, a classification network based on depth feature local resampling fusion (DFLRF) was proposed to classify and identify 36 types of rice seeds. ConvNeXt was adopted as the backbone network to extract feature representations of the seeds. Subsequently, a global feature extraction branch was constructed by using the feature intensification attention module (FIAM), while the local feature extraction branch was built by using both the multi-channel convolutional local resampling module (MCLRM) and FIAM. The global and local features were then fused, and under the supervision of the CosFace loss, rice seed varieties with highly similar visual characteristics were accurately identified. A self-collected dataset was utilized. Experimental results demonstrated that the proposed ConvNeXt-DFLRF model achieved an overall accuracy of 86.90%, marking a 5.88 percentage points improvement over the baseline model. When compared with mainstream architectures such as InceptionResNetV2 and EfficientNetV2, the proposed model exhibited improvements ranging from 2.92 to 8.8 percentage points, achieving the highest recognition performance among the tested methods. These findings confirmed that the proposed model was effective in classifying 36 rice seed varieties and provided a novel and efficient solution for multi-type rice seed classification and recognition tasks.

Abstract:Rice was selected as the research subject, and hyperspectral reflectance data of the rice canopy within the range of 400~1000nm was collected experimentally. To preprocess this data, the Savitzky-Golay smoothing method was applied, followed by the successive projections algorithm (SPA) to identify characteristic wavelengths. Based on the processed spectral data, an extreme learning machine (ELM) model optimized by the non-dominated sorting whale optimization algorithm (NSWOA) was developed to estimate the nitrogen content in the rice canopy. To evaluate its performance, the NSWOA optimized ELM model was compared with a traditional back propagation neural network (BPNN) and a standard ELM model. The results indicated that the characteristic wavelengths identified by the SPA algorithm were 400nm, 440nm, 487nm, 542nm, 589nm, 660nm, 675nm, 739nm, 766nm, 808nm, 878nm, 912nm and 949nm. The NSWOA-ELM model based on reflectance at these selected wavelengths performed best, achieving a determination coefficient (R2) of 0.8593 for the training set and 0.8543 for the validation set, with root mean square errors (RMSE) of 0.2002mg/g and 0.2069mg/g, respectively. Compared with the BPNN and standard ELM models, the NSWOA-ELM model demonstrated superior predictive accuracy and model stability. In conclusion and generalization, the NSWOA-ELM rice canopy nitrogen inversion model provided a reliable approach for assessing rice growth conditions and supporting precision fertilization strategies.

Abstract:Soil temperature is a very important variable in agricultural science, and its spatial and temporal variations are characterized by stochasticity, nonlinearity and non-stationarity, which greatly affects the accuracy of prediction, for this reason, an improved black kite algorithm (IBKA) was proposed to optimize a short-term soil temperature prediction model by integrating the attention mechanism of convolutional neural networks (ACNN) and dendrite networks (DD). Firstly, the black-winged kite algorithm was improved by adaptive hierarchical learning strategy to enhance the optimization ability of the algorithm;then the two models of convolutional neural networks (ACNN) and dendrite networks (DD) with integrated attention mechanism were fused to obtain a new model, ACNN-DD, which was used to mine the relationship between soil temperature and feature variables, and then to output the prediction of the soil temperature in the next 6 hours. Finally, in order to validate the model, soil temperature data monitored at the vegetable planting base in Fengyan Village, Fengyan Township, Nanchuan District, Chongqing, the National Field Scientific Observatory for Naiman Farmland Ecosystems, Inner Mongolia, and the National Field Scientific Observatory for Ansai Farmland Ecosystems, Shaanxi, were brought into the model. The results showed that the coefficients of determination of the model were as high as 0.98, 0.98, and 0.99, and the root mean square errors were as low as 1.12℃, 1.35℃, and 1.37℃, the mean absolute percentage errors were reduced to 3.83%, 5.54%, and 5.41%, which were all superior to that of the traditional soil temperature prediction models, such as ILSTM_Soil, MLP-FFA, and SPA-GA-SVR, etc. This showed that the model can effectively predict the soil temperature in the next 6 hours, which can provide a theoretical basis for the application in the field of smart agriculture.

Abstract:Aiming to address the issue that current grain storage condensation risk prediction mainly relies on simple grain temperature monitoring and empirical judgment, lacking accurate predictive tools, a grain storage condensation risk prediction method was proposed combining a three-dimensional convolutional neural network (3DCNN) and a temporal fusion transformer (TFT). The method utilized 3DCNN to extract spatial features of the internal temperature field of the grain heap and employed TFT to handle time series data. Based on historical meteorological data and future weather forecast data, it achieved high-precision prediction of the grain heap temperature field and used the prediction results along with the weather forecast data for the next seven days to assess condensation risk. The measured data from eight granaries in the Tacheng area of Xinjiang was used for model training and validation, achieving an accurate grain temperature prediction model. The comparative experiment on the test set showed that the 3DCNN-TFT model had a mean absolute error (MAE) of 0.16℃ and a root mean square error (RMSE) of 0.18℃ in grain temperature field prediction, significantly outperforming other predictive models. Finally, the real granary experiment validated the model’s generalization ability, with results showing that the 3DCNN-TFT model had high predictive accuracy, with an average MAE of 0.16℃ and average RMSE of 0.19℃. The model also successfully predicted condensation risk and issued early warnings. The results indicated that the proposed method significantly improved the prediction accuracy of the grain heap temperature field, accurately predicted condensation risk, and provided strong support for the development of grain storage monitoring and prediction systems.

Abstract:Due to geographical or cultural differences, the entity names in agricultural texts are confused, which makes automatic identification and extraction of information complicated and limits the development of agricultural informatization. In view of this, an agricultural entity normalization method based on mBART was proposed. Firstly, based on the knowledge and experience of experts in the agricultural field, a crop-oriented agricultural text dataset was constructed, covering the three major crops of “legumes”, “cereals” and “oil crops”, with a total of 22440 pieces of high-quality agricultural labeling data. Secondly, the problem of agricultural entity normalization involved the detection and identification of non-normalized agricultural entities. A unified generative framework was proposed based on mBART to jointly detect and identify agricultural non-normalized entities and directly complete the task of normalizing agricultural named entities. Furthermore, in order to improve the normalization effect of agricultural entities, auxiliary tasks of agricultural non-normalized entity detection and agricultural non-normalized entity recognition were additionally introduced into the model. Finally, extensive experiments were conducted on the proposed crop dataset. The results showed that the proposed method achieved P, R, and F1 above 0.99 in the task of agricultural entity normalization, and all indexes were optimal compared with other methods. Compared with the large language models, the proposed method also had significant advantages.

Abstract:The identification of individual cows is a prerequisite and foundation for realizing accurate and intelligent farming, but the identification method based on image information is easy to be affected by the environment and observation angle. In order to achieve accurate identification of cow identity under top-view conditions, an individual identification method of cow based on PointNet++ and improved ConvNeXt model was proposed. Firstly, the apex RGBD images of cows were collected, and PointNet++ model was used to locate the hook and pin bones of cows. Secondly, the curvature changes of hook and pin were analyzed to accurately locate hook and pin, and the key areas were determined according to the distance relationship between hook and pin, and the key areas were converted into two-dimensional body spot images. Finally, based on the improved ConvNeXt model, image classification was performed to achieve accurate identity recognition. A total of 6800 top view images from 30 cows were constructed, and the training set, validation set, and test set were constructed at a ratio of 7∶2∶1. The results showed that the AP50 of the point cloud segmentation model was 92.5%, and the identification accuracy of the cow can reach 94.67%. Compared with that of the original model, the classification accuracy of the improved ConvNeXt model was improved by 4.83 percentage points under the condition that the weight was basically the same. The method had high robustness to the position and angle of the cow in the top visual field.

Abstract:Aiming at the problem of weakened diagnostic effectiveness caused by disease diagnosis models relying solely on semantic feature while lacking structural feature, a diagnosis method of sheep disease was proposed based on knowledge graph and fusion of semantics and structure. Bidirectional gated recurrent unit (BiGRU) was utilized to extract semantic feature from symptom texts, and graph convolutional network (GCN) was utilized to capture latent structural feature from the symptom knowledge graph. To better integrate semantic and structural features, an improved attentional feature fusion (AFF) module was introduced. Experimental results on the sheep disease symptom dataset demonstrated that the proposed model achieved the accuracy, precision, recall, F1 score, and mean reciprocal rank of 96.86%, 97.73%, 97.32%, 97.25%, and 97.49%, respectively. Compared with models such as TextCNN, TextRCNN, TextRNN, DPCNN, LASA, HSAN-capsule, DCDKG, and CNN-BiLSTM-Attention, the accuracy was improved by at least 0.19 percentage points and up to 1.76 percentage points, the precision was improved by at least 0.18 percentage points and up to 0.84 percentage points, the recall was improved by at least 0.14 percentage points and up to 1.21 percentage points, the F1 score was improved by at least 0.21 percentage points and up to 1.50 percentage points, and the mean reciprocal rank was improved by at least 0.15 percentage points and up to 1.23 percentage points. Additionally, experiments on the public datasets Dxy and Muzhi validated the model’s robustness. To enhance the interpretability of disease diagnosis, an improved gradient-based saliency method was proposed to explain diagnostic results. The method of fusing semantic and structural features effectively improved disease diagnosis accuracy, providing technical support for sheep disease diagnosis.

Abstract:In modern animal husbandry, the breathing rate and heart rate of pigs are critical indicators for assessing their health status. Therefore, the development of a non-contact, high-precision multi-target vital sign monitoring technology holds significant importance for advancing the modernization of the livestock industry. Millimeter-wave radar technology, by transmitting linear frequency modulated continuous wave (LFMCW), achieves an extremely high pulse compression ratio, thereby significantly enhancing radar range resolution and target detection capabilities. To address the limitations of existing methods in synchronous multi-target breathing rate and heart rate monitoring, a synchronous multi-pig breathing rate and heart rate monitoring method was proposed, which integrated machine vision with millimeter-wave sensing. The YOLO v8 algorithm was employed to identify pig targets in images, effectively eliminating non-pig vibration sources and providing prior conditions for millimeter-wave radar. Subsequently, the phasor mean cancellation algorithm for LFMCW and two-dimensional Fourier transform method were utilized to remove static targets and decouple multi-target echoes. After extracting echo signals, time-frequency diagrams of breathing and cardiac activities were generated through bandpass filtering, short-time Fourier transform, and periodic evaluation metrics to calculate breathing rate and heart rates. To validate the effectiveness of the proposed method, multiple experiments were conducted in practical farm environments. Results demonstrated that the average relative error for breathing rate measurement was 4.57%, and that for heart rate measurement was 9.26%, indicating high synchronization accuracy and notable anti-interference capabilities against non-target vibration sources in the environment.

Abstract:Sex identification of pigeons is a crucial task in the breeding and pairing process. To enable intelligent sex identification of pigeons, a method based on pigeon vocalizations was proposed and the endpoint detection model EGVAN_VAD and the sex classification model EGVAN, were developed both based on the EGLKA module. The method transformed one-dimensional audio signals into two-dimensional Mel spectrograms as input to the EGVAN_VAD model for detecting pigeon vocal segments, followed by noise reduction by using a thresholding method and a Butterworth low-pass filter to eliminate both steady-state and transient environmental noise. The denoised Mel spectrogram was then used as input for the EGVAN model, which was compared with LSTM, GRU, TDNN, and VAN models for recognizing pigeon vocalizations across different age groups (1~3 months, 3~6 months, and more than six months). Experimental results indicated that the EGVAN_VAD model achieved an accuracy of 93.4% and the recall of 94.3%, with processing time of 10ms per 4-second audio segment, outperforming other endpoint detection models in overall performance. The EGVAN model achieved the highest recognition precision for female and male calls, reaching 90.7% and 90.3%, respectively, with processing time of 14.1ms per 3-second audio segment. For pigeons aged 3~6 months, the identification success rate reached 99.5%, while the average success rate across all age groups was 89.0% following systematic testing. These findings demonstrated that the EGVAN model exhibited excellent performance in recognizing both male and female pigeon calls. The research result can provide a technical reference for applying audio-based methods to intelligent and accurate sex identification in other monomorphic bird species.

Abstract:Aiming to reveal the spatio-temporal variation patterns of water consumption and economic benefits in apple production in China, the water footprint of apple production in major apple-producing provinces from 2003 to 2022 was calculated based on CLIMWAT 2.0 and CROPWAT 8.0 software. Path analysis, gravity center migration method and water footprint benefit evaluation were employed to deeply analyze the key factors influencing the spatio-temporal variation of water footprint in apple production and its economic benefits. The results showed that the water footprint of apple production in major apple-producing provinces in China was increased first and then decreased over time, and presented a pattern of “high in the west and low in the east” in space. The average annual water footprint of apple production in major apple-producing provinces was 1.41×1010m3, among which blue, green and grey water footprints accounted for 20.5%, 52.1% and 27.4%, respectively, with green water footprint being the dominant factor. Further analysis indicated that the amount of pesticide used, total power of agricultural machinery and agricultural water consumption were the three core factors affecting the total water footprint of apple production, and total power of agricultural machinery, relative humidity and the amount of pesticide used had positive driving effects on water footprint. In addition, the output value of water footprint showed a downward trend, indicating that the economic benefits of water footprint in apple production needed to be improved. Based on this, it was suggested that the major apple-producing provinces should optimize the planting layout and structure, promote water-saving irrigation technologies, improve water resource utilization efficiency, and promote the green and sustainable development of the apple industry.

Abstract:Climate change will significantly alter the water and heat resources necessary for cultivating water-saving and drought-resistant rice, thereby affecting the suitable areas for such cultivation. A geographic information system (GIS) and a multi-criteria decision-making (MCDM) model (BCM-GIS) were established. The impacts of climate, topography, and soil were comprehensively considered to evaluate the suitability area for drought-resistant and water-saving rice in the Southwest China under both currents (2000—2019) and future climate change scenarios (RCP4.5 and RCP8.5). The results indicated that the BCM-GIS model had a high prediction accuracy with an area under curve (AUC) value of 0.884, confirming its applicability for suitability analysis of water-saving and drought-resistant rice cultivation areas. Annual rainfall, equal to or greater than 10℃ cumulative temperature during the reproductive period, slope, and soil texture had significant effects on the distribution of suitable cultivation of water-saving and drought-resistant rice. The current highly suitable zones for water-saving and drought-resistant rice were primarily located in Dehong, Lincang, Pu’er, Xishuangbanna, Wenshan, and Qujing in Yunnan Province;Qianxi, Anshun, Qiannan, Tongren, and Qiandongnan in Guizhou Province;Chengdu, Neijiang, Zigong, Dazhou, and Guang’an in Sichuan Province;and the central-western regions of Chongqing City. These areas accounted for 20.82% of the total cultivated land in Southwest China, approximately 3.279×104km2. The medium-suitability and low-suitability zones accounted for 43.83% and 35.25% of the total cultivated land in Southwest China, respectively, amounting to approximately 6.902×104km2 and 5.552×104km2. Under the future RCP4.5 and RCP8.5 climate scenarios, the areas of water-saving and drought-resistant rice in the highly suitable zone were increased significantly during 2020—2099 and showed a northward trend, while the areas of the medium suitable zone and the low suitable zone decreased to some extent. The latest multi-criteria decision-making methods were used and the response of suitable areas for water-saving and drought-resistant rice were analyzed.

Abstract:Aiming to investigate the effects of different CO2 fertilization methods on photosynthesis and yield of grapes in facilities, the Eurasian grape variety ‘Flame seedless’ was used as the experimental object, combined with aerated irrigation technology and CO2 injection system, a closed artificial chamber was used to set up three different CO2 fertilization methods: CO2 aerated irrigation fertilization (combined with underground drip irrigation pipeline CO2 application, IW), CO2 traditional fertilization (using chemical reaction bag CO2 application, TW), and no fertilization (blank control group, CK). The photosynthetic pigment content of grape leaves, photosynthetic physiological characteristics, leaf enzyme activity, and fruit yield under different CO2 fertilization methods were measured. The results showed that within the same irrigation cycle, both CO2 fertilization methods significantly increased the photosynthetic pigment content, ratio of chlorophyll a content to chlorophyll b content (chlorophyll a/b), net photosynthetic rate, and water use efficiency of greenhouse grape leaves. Among them, the efficiency of CO2 aerated irrigation fertilization treatment was significantly higher than that of traditional fertilization treatment. CO2 aerated irrigation and fertilization treatment significantly reduced stomatal conductance and transpiration rate within the same irrigation cycle. The light saturation point and maximum net photosynthetic rate of grapes treated with CO2 aerated irrigation and fertilization were 18.4% and 21.0% higher than those of the CK control, respectively, while the light compensation point was decreased by 13.3%. In addition, CO2 aerated irrigation and fertilization treatment significantly increased leaf CO2 saturation point, CO2 compensation point, superoxide dismutase (SOD), catalase (CAT), peroxidase (POD) and ribulose diphosphate carboxylase (Rubisco) enzyme activities, and the highest grape yield was 13875.61kg/hm2. In summary, the CO2 aerated irrigation and fertilization method significantly improved the photosynthetic pigment content and photosynthetic performance of greenhouse grape leaves, improved grape yield and quality, and had the best effect. This research result can provide a certain theoretical basis for the application of CO2 fertilization technology in future facility grape production.

Abstract:In order to reveal the influence mechanism of soil water movement in typical sand-interlayered soil under different irrigation systems in Hetao Irrigation District (HID), HYDRUS-1D model was used to simulate the process of soil water movement in typical sand-interlayered soil, and the effects of different sand depth and irrigation regimes on soil water dynamics were studied based on the field experiments with three irrigation levels of high water (W1), medium water (W2) and low water (W3) , high nitrogen (F1) , medium nitrogen (F2) and low nitrogen (F3) were set in the farmland at sand depth more than 40cm. The results showed that the deeper the depth of sand layer was, the more effective it was to reduce the loss of soil water leakage and improve the root water uptake above the sand layer. With the increase of depths of sand layer and irrigation amount, the effect of reducing soil water leakage was more obvious. When the depth of sand layer was 20~40cm, the soil water leakage at the upper boundary of sand layer accounted for 12.39%~98.19% of irrigation amount. When the depth of sand layer was greater than 40cm, the soil water leakage was decreased to 2.73%~7.00% of irrigation amount. The maximum water stress of roots occurred between the second irrigation and the third irrigation, accounting for 56.04% ~68.56% of the total stress, and the water stress from jointing to the first irrigation was the smallest. The total stress on root was decreased with the increase of sand depth. Compared with the stress on 20~40cm, the total water stress of root was decreased by 10.76%~45.63%. The maximum stress on roots at different growth stages was on the depth of 20~40cm, and the stress was decreased gradually with the increase of sand layer depth, the maximum decrease was 82.61%. Considering the soil water movement and root stress in the sand layer, it was suggested that the irrigation scheme of spring maize on the typical sand-interlayered soil should be set due to the distribution of sand layer. The recommended irrigation regime for different depths of sand layer was as follows:189~216mm for the depth of 20~40cm, 243~270mm for the depth of 40~60cm, and 216~270mm for the depth greater than 60cm. The research results can provide theoretical guidance for the formulation of irrigation system on typical sand-interlayered soil in HID.

Abstract:Mechanical compaction is one of the key factors leading to changes in soil structure and function, thereby constraining the sustainable development of agriculture.Three different types of soil, i.e. sand, loam and clay, were collected. Mechanical compaction was applied to simulate various compaction levels of soil samples, and the samples were subsequently scanned by using an advanced Skyscan 1275 Micro-CT system. The performance of three different algorithms—active contour (AC), marker based watershed (MBW) and K-means—on the segmentation of Micro-CT images of soil pores and aggregate structure was compared and the K-means image segmentation algorithm exhibited the best performance. Furthermore, the morphological watershed algorithm and Canny edge detection algorithm were used to effectively achieve the in-situ separation of individual soil aggregates and pores. Based on the K-means segmentation algorithm, a 3D in-situ visualization approach was established to quantitatively characterize the physical structure of soil using Micro-CT, enabling the visualization of pore and aggregate structures in different compacted soil samples. The Micro-CT imaging results were in excellent agreement with laboratory analyses. These findings can provide a rapid and environmentally friendly analytical tools for evaluating soil compaction and digitally mapping soil structure.

Abstract:In order to obtain the optimal composting process parameters of organic fertilizer decomposition quality and efficiency, Hertz-Mindlin with JKR was used as the discrete element contact model between the material and the equipment, and according to the simulation results of particle distribution, the material mixing during tank rotation was simulated and calculated by EDEM 2022 R1 software, and the difference of material mixing degree under different tank turns in a single pile turning was determined. The dispersion coefficient was used to evaluate the mixing uniformity. The test showed that when the fermentation tank was transferred to the tank for 2r, the material was basically mixed evenly, so the frequency of a single turning tank was set to 2r, the mixture of chicken manure and corn straw was selected as the raw material for the turnover frequency test, the temperature and moisture content of the fermentation material under different turning frequencies were analyzed, and the moisture content, temperature and maturity of the fermentation material were analyzed in four kinds of phased frequency conversion mode, and the results showed that mode 4 (no turning during the heating period, turning the pile once a day during the high temperature period, and turning the pile once a day during the holding period), the highest fermentation temperature was 68.4℃, and the germination index of radish seeds treated with organic fertilizer was 107.6%, which met the requirements of decomposition and the fermentation process was stable and reliable. The research results can provide equipment and process support for the preparation of bio-organic fertilizer, ensure the decomposition quality of organic fertilizer, and effectively solve the problem of unclear fermentation process.

Abstract:Rapeseed is the most important oil crop in China. In production practices, combined harvesters are commonly employed for one-time harvesting, resulting in significant variations in chlorophyll content among harvested rapeseeds. These variations adversely affect oil quality and increase processing costs, making real-time, on-site chlorophyll content detection crucial during rapeseed procurement and processing. To develop a portable, low-cost, and non-destructive rapid chlorophyll detection device for rapeseed, an insertion-type probe was specifically designed based on the storage characteristics of bulk rapeseed. A low-cost, non-destructive chlorophyll detection device was developed using a multispectral sensor. A partial least squares (PLS) prediction model for rapeseed chlorophyll content was established based on the developed device. After preprocessing with spectral shape feature (SSF) and successive projections algorithm (SPA), the model achieved a coefficient of determination (R2) of 0.9414 and a root mean square error (RMSE) of 2.0261mg/kg. Additionally, a real-time analysis and control software was developed by using C++ on Raspberry Pi 3B with QT5, enabling one-click human-machine interaction. Finally, the accuracy and stability of the device were externally validated by using 25 rapeseed samples not involved in modeling. Each sample was measured three times, yielding an average coefficient of variation (CV) of 3.27% for chlorophyll content. The predicted values exhibited an R2 of 0.9358 and an RMSE of 1.9071mg/kg compared with that of reference physicochemical measurements. The results demonstrated that the insertion-type non-destructive chlorophyll detection device met the requirements for on-site real-time detection, providing technical support for the industrialization of rapeseed processing.

Abstract:Fishmeal, a high-value and necessary protein feed raw material, confronts difficulties in assessing authenticity. The conventional identification process is rendered challenging due to the sophistication of visual discrimination and the inefficacy of microscopic testing techniques by inspector, particularly in the presence of subtle and multiple adulterants. These issues were tackled by proposing an end-to-end segmentation model based on the Masked-attention Mask Transformer (Mask2Former) architecture. Focused on identifying multi-target adulteration feature in complex fishmeal backgrounds under high-resolution conditions to realize automatic recognition and segmentation. Fishmeal was used and adulterated with animal-derived meat powder as an example, an adulterated fishmeal microscopic image dataset was created and the adulteration characteristics were classified into five types of tissues: abnormal muscle, bone, skin, blood, and hair based on morphological analysis. Then an interactive annotation program that utilized color threshold similarity to provide pixel-by-pixel suggestive annotations was developed. The updated Mask2Former model integrated ResNet50, multi-head Attention mechanisms, and multi-scale feature processing mechanisms to extract global information and multi-resolution features. Bidirectional feature pyramid network (BiFPN) enhanced the pixel decoder, thereby improving the model’s performance in processing small-scale targets. The incorporation of the Masked Attention module limited the scope of cross-attention calculations, enabling the model to more effectively focus on the target areas within the mask. The implementation of learnable scale-level embeddings and the reordering of self-attention and cross-attention within the decoder section bolstered the model’s feature learning capabilities. Furthermore, by adding dropout operations after the multi-head Attention layer, the improved Fishmeal-Mask2Former model achieved the overall accuracy rate of 98.56% in distinguishing genuine or adulterated samples. In the fine segmentation stage, the mean average precision of abnormal adulteration feature recognition reached 80.52 %, the mean average recall reached 76.01%, the mean F1 score reached 78.86%, and the segmentation mean accuracy was improved to 82.08%. The method proposed had significant advantages over traditional methods. Finally, a visual strategy with its interface of the segmentation results was designed, aimed to provide an intuitive, accurate, and efficient microscopic visual automated identification method for the fishmeal adulteration in its quality detection.

Abstract:In order to solve the problem of poor adaptability of energy management strategy for hydrogen fuel cell electric tractor (HFCET) when running online, a hybrid energy management strategy was proposed based on deep Q-networks (DQN) learning. Applying deep reinforcement learning method to hydrogen fuel cell (HFC) electric tractor played an important role in improving the fuel cell economy and prolonging the service life of fuel cell. Firstly, Q-learning method was compared with deep Q-networks learning method and dynamic programming (DP) method with the fuel cell hydrogen consumption as the target. Secondly, the fuel cell performance degradation factor was incorporated into the objective function, and the dynamic balance between the hydrogen fuel cell economy and system performance degradation was achieved by adjusting the performance degradation factor and hydrogen consumption. Finally, the effectiveness of the proposed strategy was verified by the actual operating condition of the electric tractor. The experimental results of actual operating conditions showed that when hydrogen fuel cell performance degradation factor was included in the training, the energy consumption of energy management strategy (EMS) was decreased by 2.46%, reaching 87.63% of the actual operating condition of dynamic programming method energy management strategy, which effectively inhibited the decline of hydrogen fuel cell performance. At the same time, compared with dynamic programming method, the computational efficiency was increased more than 78%.

Abstract:The optimization of structural parameters had a significant impact on the performance of mechanisms, in this paper, a five-degree-of-freedom hybrid mechanism for complex surface machining was taken as the research object, and the kinematic analysis and optimization of structural parameters were carried out. The forward and inverse kinematic models of the mechanism were established respectively by using the numerical iterative method and the analytical method. The layered search method was used to solve the reachable workspace and calculated the volume of the regular workspace consisting of the maximum internal tangent circle of each layer. The global performance index was defined and the influence of the size parameters of the mechanism on each performance index was analyzed by orthogonal test method. The multi-objective optimization model of the mechanism was established by the reference target distance method and the linear weighting method, and the genetic algorithm was used to optimize the design. The simulation results showed that the size of the optimized mechanism was reduced, while the global motion performance index, the global motion performance fluctuation index and the global operability performance index were increased by 40%, 45% and 77% respectively. To further verify the optimization results, the velocity and acceleration of the models before and after optimization were simulated and compared. It was found that after the mechanism was optimized, the impact and vibration during the motion process could be reduced, making the motion control of the mechanism more precise.