This study explores the potential of harvesting and utilizing thermoelectric energy generated during biochar production. Biochar, a carbon-rich material produced from biomass pyrolysis, offers a sustainable pathway for waste management and soil improvement. However, the biochar production process also generates significant heat, much of which remains underutilized. This research investigates the integration of thermoelectric generators (TEGs) into biochar production equipment to convert waste heat into electricity. The system's design and optimization aim to enhance energy efficiency while supporting renewable energy goals, by using the harvested energy to power up small devices, such as sensors. Experimental results demonstrate the feasibility of capturing thermal energy and converting it into usable electrical power, contributing to the overall sustainability of the process. The experimental results revealed a maximum output voltage of 1.226 V, recorded at a temperature gradient of approximately 42.0 °C, a peak power output of approximately 37.6 mW, an overall maximum temperature gradient of approximately 42.8 °C reached slightly before the voltage peak, and a cumulative harvested energy of 288.92 J. The findings also highlight key factors influencing thermoelectric conversion efficiency, including temperature gradients, TEG material properties, and system integration challenges. This approach not only mitigates energy losses but also aligns with global efforts toward carbon neutrality and resource optimization. The study underscores the dual benefit of biochar production: providing agricultural benefits while supporting energy sustainability.
The seedling stage is a relatively short but critical period in the maize growth cycle, during which rapid weeding, pesticide application, and fertilization are required. Crop row centreline extraction technology based on computer vision can assist agricultural machinery in performing automated field operations, thereby improving work efficiency. Therefore, improving the accuracy of crop row centreline extraction imposes higher requirements on existing models and algorithms. In this study, a model named Synergistic Attention YOLO (SA-YOLO) was proposed. The feature extraction capability for crop rows was enhanced by introducing the spatial and channel synergistic attention module and designing the A2C2f module integrated with switchable atrous convolution. In addition, a crop topological skeleton feature point extraction strategy and a collaborative fitting method combining principal component analysis and least squares was integrated to accurately extract crop row centrelines from the skeleton feature points of crop pixels within the segmented region of interest. Comparative experiments showed that the SA-YOLO model achieved competitive performance, with an average precision at IoU threshold 0.5 (AP50) of 99.1%, and a mask Intersection over Union (IoUmask) of 83.8%, while maintaining a lightweight architecture with only 3.72 M parameters. The mean angle error (𝐴𝑚𝑒𝑎𝑛) of crop row centerline extraction was 0.590°, and the mean normalized lateral error (𝐿𝑚𝑒𝑎𝑛𝑛𝑜𝑟𝑚) was 0.357%. This study provides a new approach for exploring crop row centerline extraction in agriculture and can further enrich the theoretical and technical foundation for visual navigation of agricultural robots
The quality inspection of sorghum grains after drying is critical for ensuring storage stability and processing quality. However, traditional manual inspection methods are inefficient, time-consuming, and prone to subjective biases. To address these challenges, this study proposes an improved YOLOv8s model, named YOLOv8s-GBD, for accurate sorghum grain classification and detection. The proposed model integrates three key enhancements. First, the DFC Attention-based bottleneck structure from GhostNetV2 is incorporated into the C2f module to enhance global context awareness and improve detection accuracy. Second, a bi-directional feature pyramid network (BiFPN) is introduced to optimize multi-scale feature fusion. Third, a dynamic head framework based on attention mechanisms is adopted to strengthen feature representation, further boosting the model’s accuracy and robustness. Experimental results demonstrate that the YOLOv8-GBD model achieves 95.2% precision, 98.0% recall, 98.7% mAP1, 97.5% mAP2, and 96.6% F1-score. Compared to the original YOLOv8s model, these metrics show improvements of 3.1%, 1.7%, 1.2%, 1.0%, and 2.4%, respectively. Furthermore, the YOLOv8-GBD model outperforms other YOLO series models in sorghum grain classification and detection tasks. In conclusion, the YOLOv8s-GBD model meets the requirements for high accuracy in sorghum grain quality inspection, offering a robust solution for practical applications.
This study compares energy and exergy performance of two vapour-compression refrigeration systems designed to provide simultaneous refrigeration and freezing at two distinct temperature levels for food-industry applications. The simulation results indicated that R1234yf provided the best overall performance for both configurations. The exergy analysis showed that the throttling processes in the first configuration had the fastest increase in exergy destruction with the variation of ambient temperature. The structural improvement of the second configuration led to an increase in overall system performance, confirming the importance of refrigerant selection and structural optimisation of the refrigeration system. For R1234yf, the energy coefficient of performance increased from 2.69 to 2.97 at an ambient temperature of 35 °C and from 1.98 to 2.24 at 45 °C when configuration (b) was used instead of configuration (a).
Temporary grasslands are recognized both for their ecological functions, such as nutrient cycling, soil quality, and also for their potential role within crop rotation frameworks. While many studies have shown the agronomic and environmental benefits of permanent grasslands, fewer have analysed the impact of temporary grassland phases as precursors, influencing coming crops in terms of system stability or yield. Results are often mixed or contradictory in the academic literature, due to methodological limitations or insufficient statistical power. The current long-term study (2008-2014) investigates the role of grassland as a precursor to suitable arable crops (potatoes, spring wheat, silage maize and forage turnips), on a rotation-based empiric design, focusing on temporal variability, yield fluctuations and the broader agronomical context in which these changes occur. Using multi-year data, we look at crop yield performance, and discuss results offering suggestions for future research. To account for structural yield differences among crops, yields were standardized within crop categories, allowing for a unified analytical framework. Results from factorial ANOVA models indicate that grassland phases contribute to yield stabilization rather than a uniform increase in production of other crops, with outcomes influenced by crop type and stationary area conditions. Our results suggest that grasslands have the potential to enhance system resilience and are valuable for sustainable crop production under increased environmental uncertainty.
Precision agriculture requires advanced digital solutions capable of continuously monitoring environmental conditions and pest dynamics in orchard ecosystems. This paper presents the development, integration, and experimental validation of a digital system dedicated to orchard crop monitoring. The proposed system integrates a smart pheromone trap equipped with an HD camera, a meteorological monitoring station, an autonomous photovoltaic power supply system, and a wireless communication module for real-time data transmission. The experimental model was designed to support automatic pest detection and continuous acquisition of climatic parameters associated with pest and pathogen development. The monitoring architecture combines environmental sensing, image acquisition, local data processing, and cloud-based communication. Functional testing was conducted under controlled orchard conditions over a 30-day monitoring period. The obtained results demonstrated stable autonomous operation, reliable wireless communication, and accurate environmental monitoring. The average deviations compared with a reference meteorological station were within ±0.4 °C for air temperature, ±2.5% for relative humidity, and ±0.3 m/s for wind speed. The smart trap achieved pest identification accuracy higher than 90% for target species. The photovoltaic system ensured autonomous operation for more than 72 h without direct solar radiation. The developed digital platform represents a scalable solution for integrated pest management and precision orchard monitoring applications.
Post-mining substrates are spatially heterogeneous and require site-specific diagnosis before biomass cultivation. This study documents the establishment and three-year monitoring of a 4,500 m² field platform on a coal-mining waste dump in Petrila, Romania. Nine 500 m² plots were used: three fixed-crop reference plots (designated as controls in the original project records) and six plots under a balanced cyclic rotation of maize (Zea mays), soybean (Glycine max) and sweet sorghum (Sorghum bicolor). Amendment treatments formed part of the wider experimental layout but are not evaluated in the present manuscript. Baseline measurements at eight points in June 2023 included instrument-reported volumetric water content (VWC), temperature, penetration resistance, field pH and semi-quantitative N-P-K screening, complemented by preliminary potentially toxic element screening. Mean VWC was 26.6 ± 4.2% at 5 cm and 20.0 ± 4.3% at 15 cm; mean temperatures were 18.2 ± 1.3 and 16.7 ± 1.1°C. Mean penetration resistance across seven complete profiles was 1.05 ± 0.28 MPa and increased with depth. Field pH averaged 7.1 ± 0.1, and the available laboratory summary prioritised Cu and Ni for follow-up. All crops reached reproductive development. Preliminary operational harvest totals were 85, 263 and 487 kg in 2023, 2024 and 2025, respectively; these values are reported descriptively only and are not interpreted as crop yield because harvested area, crop-specific contributions and moisture basis were not documented consistently. The nine crop-year elemental records are likewise treated as single descriptive observations displaying apparent crop-year differences in the reported Cu, Ni, Zn and Pb values; no inferential comparison among crops or years and no conclusion on soil-to-plant transfer or bioaccumulation is made.
To address the problems of low manual pickup efficiency and severe soil accumulation in existing shovel-type pickup devices used for segmented potato harvesting, a composite pickup device consisting of a grid-type segmented pickup shovel and a flexible lifting wheel was designed. A potato-soil-machine discrete element simulation model was established using EDEM to optimize structural parameters, including grid spacing, and to investigate the effects of operating parameters, including forward speed, digging depth, and lifting-wheel speed, on pickup performance. The simulation results showed that grid spacing was negatively correlated with the loss rate and positively correlated with the damage rate. An optimized grid spacing of 25 mm reduced the damage rate to 2.21% and the loss rate to 3.70%, while effectively reducing potato jamming. As the forward speed and lifting-wheel speed increased, the loss rate decreased, whereas the damage rate increased. Digging depth had a nonlinear effect on the loss rate, which initially decreased and then increased, reaching its minimum at a digging depth of 100 mm, whereas the damage rate continuously decreased with increasing digging depth. Field tests yielded an average loss rate of 3.88% and an average damage rate of 2.24%, meeting the operational requirements for potato harvesting.
Airborne dust generated during straw pickup can impair operating visibility and local air quality; however, the respective contributions of mechanical particle detachment and aerodynamic entrainment in flail-type pickup headers remain insufficiently quantified. This study combined discrete element method (DEM) and computational fluid dynamics (CFD) analyses with field measurements to investigate these two processes. A bonded corn root stubble–soil DEM model was established in EDEM, and the bond fracture rate was used to characterize the relative degree of particle detachment. The effects of ground clearance (30–90 mm), rotor speed (4100–5300 rpm), forward speed (0.5–1.5 m/s), plant spacing (200–300 mm), and row spacing (300–500 mm) were evaluated. CFD simulations were then used to characterize the time-averaged airflow field within the pickup-header housing. Increasing the ground clearance from 30 to 90 mm reduced the bond fracture rate by 44.5%, whereas increasing the rotor speed from 4100 to 5300 rpm increased it by 52.3%. By relating the DEM particle-detachment states to the CFD flow-field characteristics, three dust-generation pathways were identified: stubble fragmentation, root extraction, and direct entrainment of loose surface particles. Rotor speed affected both mechanical detachment and airflow entrainment at the header inlet. Field measurements showed that the selected low-dust operating condition, with a ground clearance of 75 mm, rotor speed of 4400 rpm, and forward speed of 0.75 m/s, reduced total suspended particulate (TSP), PM10, and PM2.5 concentrations by 62.2%, 58.4%, and 53.8%, respectively, compared with the dust-intensive condition. By linking particle detachment, airflow entrainment, and field measurements within a mechanism-based framework, this study clarifies the effects of operating parameters on dust generation and provides a basis for selecting low-dust operating parameters and improving pickup-header inlet design.
To address the high penetration resistance and severe soil adhesion of tobacco planting devices operating in wet, clayey soil, a biomimetic planting device was designed using a Gaussian multi-peak function and characteristic curves derived from multiple bird beaks. The peak parameters, including height, position, and width, were optimized using discrete element method (DEM) simulations and response surface methodology. The resulting biomimetic profile reduced penetration resistance and soil adhesion, thereby improving tobacco transplanting performance. The optimized prototype exhibited a penetration resistance of 18.7 N and an adhered soil mass of 14.98 g, representing reductions of 60.47% and 40.56%, respectively, compared with the conventional prototype.
A study was carried out to evaluate the performance of mobile and stationary fish smoking technologies. The experiment was conducted using a randomized block factorial design, with smoking temperature as a blocking factor. According to the investigation findings, at a smoking temperature of 90℃ and a smoking time of 4 hours, the maximum moisture content reduction of 81.4% was noted. In contrast, at a smoking temperature of 70 ℃ and a smoking time of 2 hours, the minimum moisture content reduction of 57.6% was noted for the electric fish smoker. Based on the test results of the electric fish smoker, at 90 ℃ of smoking temperature and 4 h of smoking time, the maximum drying rate of 0.69 kg h-1 was obtained. The findings verified that the maximum thermal efficiency of 69.8% was recorded at a smoking temperature of 90 ℃, and a smoking time of 2 h for the electric fish smoker. Based on comparative evaluation results, the moisture content reduction, drying rate, fuel consumption and thermal efficiency for the fuelwood fish smoker were found to be 66.17%, 0.38 kg h-1, 4.8 kg kg-1 fish, and 37.9%, respectively. The study demonstrated that electric fish smoking technology significantly outperformed fuelwood smoking technology across all evaluated parameters. This study revealed that the electric fish smoking technology is strongly recommended for demonstration and scaling up to end users, particularly in areas where traditional fish smoking methods remain dominant.
Illegal dumping in peri-urban and natural environments poses persistent challenges for environmental authorities, particularly when large areas must be inspected with limited personnel and budget. Reliable and scalable monitoring approaches are essential for practical environmental management. Unmanned Aerial Vehicles (UAVs) offer rapid spatial coverage, yet it remains unclear how complex the analysis must be to reliably detect waste sites in practice. This controlled pilot study evaluated the operational reliability of three UAV-based image analysis approaches (global statistical screening, classical change detection with colour filtering, and deep learning object recognition) under controlled field conditions. The field experiment used controlled contamination conditions so that the evaluation reflected practical detection reliability rather than algorithm benchmarking. The statistical method could not reliably identify localised waste, while change detection achieved moderate reliability (~72%; 29 of 40 evaluated acquisitions) but required reference imagery. The deep learning model achieved the highest site detection reliability (~93%; 37 of 40 evaluated acquisitions) and worked without baseline images. Its main advantage was improved object completeness rather than better identification of contaminated areas. These observations showed that monitoring workflow design matters more than increasing algorithmic complexity. A tiered strategy that combines broad screening with focused high-accuracy analysis can reduce inspection effort without compromising the results of inspection. The evaluated pilot framework supported the selection of monitoring approaches according to available resources, and the findings apply to the controlled, low-altitude field conditions (15 m above ground level, single test site) examined here rather than to fully validated multi-site operation.
To address parameter optimization and fertilization uniformity of a fluted-wheel fertilizer metering device for granular organic fertilizer, a 3D-scanned fertilizer model was constructed, and an EDEM fertilization simulation model was established. The Poisson’s ratio and elastic modulus of the fertilizer were determined using a texture analyzer. Drop tests were conducted to determine the coefficients of restitution between fertilizer particles and between the fertilizer and PA, PC, and PLA. Inclined-plane rolling tests were conducted to determine the rolling friction coefficients between the fertilizer and PA, PC, and PLA. A cylinder-lifting test for measuring the angle of repose, combined with EDEM simulations and a Design-Expert experimental design, was used to determine the static and rolling friction coefficients between fertilizer particles, as well as the static friction coefficients between the fertilizer and PA, PC, and PLA. By combining the discrete element method (DEM) with response surface methodology, the effects of the parameters on fertilization uniformity were ranked as follows: fluted-wheel inscribed-circle diameter > fluted-wheel lead angle > number of grooves; fertilizer-metering shaft rotational speed > fertilizer-discharge flap angle > effective working length of the fluted wheel. The optimal parameter combination was determined to be a fluted-wheel inscribed-circle diameter of 37 mm, 7 grooves, a fluted-wheel lead angle of 63°, an effective working length of the fluted wheel of 31 mm, a fertilizer-metering shaft rotational speed of 28 rpm, and a fertilizer-discharge flap angle of −1°. Soil-bin tests using 3D-printed fluted wheels with different structural parameters produced results that were generally consistent with the simulation results, providing a reference for the DEM-based optimization and design of fluted-wheel fertilizer metering devices.
The key challenge in Brassica napus combine harvesting is achieving a balance between threshing rate and impurity rate, both of which are strongly influenced by stem breakage. This study investigated the breakage characteristics and static mechanical properties of Brassica napus stems from the middle and lower reaches of the Yangtze River using axial tensile tests combined with digital image correlation (DIC). The results showed that the fiber layer area of rapeseed stems increased with stem diameter, and stem breaking force was significantly positively correlated with stem diameter. For stem diameters ranging from 7 to 16 mm, the breaking force ranged from 166.82 to 970.82 N and was described by the relationship Fb=42.064dx1-92.894. The tensile strength of rapeseed stems ranged from 25.1 to 64.6 MPa, and a quadratic response surface model relating tensile strength to stem diameter and moisture content was established. DIC measurements of strain in the radial and axial orthogonal directions showed that axial deformation was substantially smaller than radial deformation, with an average axial strain of εy=4.22±2.23×10-3 and an average radial strain of εx=3.84±2.17×10-2. These findings improve the understanding of the mechanical response of rapeseed stems under tensile loading and provide a theoretical basis for optimizing threshing devices for efficient, low-damage rapeseed harvesting.
Seedling replenishment during the operation of high-speed ride-on rice transplanters still relies heavily on manual labor, thereby limiting the level of operational automation. To address this limitation, an automatic seedling-loading system comprising a longitudinal conveying mechanism, a lateral drive mechanism, and a tracking compensation mechanism was developed. Based on an STM32 embedded platform, a control strategy for detecting seedling shortages and performing precise seedling-loading operations was developed, enabling automatic shortage detection, lateral positioning, and targeted seedling replenishment. The main operating parameters—conveyor belt speed, lateral offset distance, and fore-aft tilt angle—were first investigated individually to determine their appropriate ranges. Subsequently, a second-order, three-factor, five-level rotatable orthogonal experiment was conducted. Seedling-loading success rate was used as the evaluation index, and a multifactor regression model was established to analyze the interactions among the factors. Bench tests were conducted to optimize the system parameters and determine the optimal parameter combination. Subsequent field trials confirmed the operational reliability of the system, with an average seedling-loading success rate of 90.67% over five batches of tests. The proposed system shows considerable potential for replacing manual loading of mat-type rice seedlings and provides a practical solution for advancing the full-process automation of rice transplanting operations.
Egg appearance defect detection in conveyor-belt environments is challenged by irregular defect morphology, illumination variations, and the limited computational resources of intelligent sorting equipment, making it difficult to simultaneously achieve high detection accuracy and lightweight deployment. In this study, a lightweight detection model, termed YOLOv11n-DAL, is proposed to improve egg defect detection performance while reducing computational complexity. The proposed model integrates C3k2_DCNv4, ADown, and Detect_LSDECD modules into the YOLOv11n framework to enhance defect feature representation, improve effective information retention during downsampling, and reduce computational redundancy in the prediction stage. A conveyor-belt egg image dataset containing three categories, namely damaged, dirty, and good, was constructed. The dataset consisted of 1,730 original images, and the training set was expanded to 4,152 images through data augmentation. Comparative experiments and ablation studies were conducted to validate the effectiveness of the proposed model. The experimental results demonstrated that YOLOv11n-DAL achieved 95.2% Precision, 94.5% Recall, 94.8% F1-score, 96.7% mAP@0.5, and 95.8% mAP@0.5:0.95. Compared with the original YOLOv11n model, YOLOv11n-DAL reduced the number of parameters from 2.6 M to 1.7 M, GFLOPs from 6.3 to 4.8, and model size from 5.2 MB to 3.8 MB, while maintaining superior detection performance. The results indicate that YOLOv11n-DAL achieves an effective balance between detection accuracy and computational efficiency, providing a lightweight and practical solution for intelligent egg sorting systems in conveyor-belt environments.
To investigate the concentration of molasses ethanol wastewater, a two-dimensional numerical model of falling-film evaporation over horizontal tubes was developed. The model used the Volume of Fluid (VOF) method to capture the liquid-gas interface and was coupled with the Lee phase-change model. The model was validated by comparing the predicted liquid-film thickness and average heat-transfer coefficient with experimental data reported in the literature. The validated model was then used to systematically investigate the effects of spray density (0.10 ~ 0.30 kg/(m·s)) and heat-transfer temperature difference (2 ~ 8 °C) on the falling-film flow and evaporation performance of molasses ethanol wastewater. The results showed that the liquid-film thickness initially decreased and subsequently increased along the circumferential direction, reaching a minimum within the angular range of 90° ~ 120°. Increasing the spray density produced a thicker liquid film; nevertheless, the average heat-transfer coefficient increased monotonically by 23.0% ~ 27.2%, whereas the mass-transfer rate decreased by 19.3% ~ 25.1%. Increasing the heat-transfer temperature difference increased both the average heat-transfer coefficient and the mass-transfer rate. When the temperature difference increased from 2 to 8 °C, the mass-transfer rate increased 2.84 ~ 3.07-fold, and its effect on evaporation performance was substantially greater than that of spray density. Therefore, where operating conditions permit, a lower spray density combined with a higher heat-transfer temperature difference is recommended for the falling-film evaporative concentration of molasses ethanol wastewater.
This study develops an artificial neural network (ANN) model to optimize the convective drying process of pear slices. Experimental data from drying at 50°C, 60°C, and 70°C were used to train an ANN with three hidden layers. The model predicted drying rate with high accuracy (98.823% validation fidelity), capturing the nonlinear relationship between moisture content, time, and drying rate. Results demonstrated that higher temperatures accelerated drying but required careful control to maintain quality. The ANN effectively identified optimal drying parameters, balancing energy efficiency and product quality preservation, providing a valuable tool for industrial drying process optimization.
Winter potato cultivation in paddy fields in southern China can improve land-use efficiency and increase farmers’ incomes. However, compared with those in northern China, paddy soils in southern China are generally heavy, cohesive, and compacted, resulting in poor mechanized potato-harvesting performance and restricting the development of the potato industry. To improve potato-soil separation efficiency under heavy clay soil conditions, a twin-roll soil-crushing device was installed ahead of a conventional chain-and-rod potato-soil separation device based on the characteristics of clayey soils in southern China. This configuration formed a two-stage separation process involving soil crushing followed by screening. The soil-crushing process was theoretically analyzed, a discrete element model of the soil-crushing and screening device was established, and the soil-crushing and potato–soil separation processes were numerically simulated. Response surface simulation experiments were conducted using rib height, number of ribs, and twin-roll rotational speed as the experimental factors, with soil-screening efficiency and the maximum force acting on the potato tubers as the response variables. The results showed that the twin-roll soil-crushing device effectively disturbed and crushed the soil during operation. Increasing the rib height and twin-roll rotational speed enhanced the soil-disturbance and crushing effects. The optimal structural and operating parameters of the twin-roll soil-crushing device were a rib height of 30 mm, three ribs, and a twin-roll rotational speed of 323.70 r/min. A simulation verification test conducted using these optimal parameters yielded a soil-screening efficiency of 72.80% and a maximum force acting on the potato tubers of 15.45 N. Field-test results showed that the potato harvester equipped with the soil-crushing device achieved a tuber damage rate of 0.285%, a skin-damage rate of 2.38%, and a tuber loss rate of 3.18%, thereby meeting the requirements of the applicable national standards. These findings provide a reference and technical support for the development of potato-harvesting machinery suitable for heavy clay soil conditions.
This study focuses on the development of a multifunctional agricultural robotic platform. Previous articles in this series established the need for such a platform based on the specific soil and climatic conditions of the Republic of Armenia. Tests of a laboratory prototype indicated the need for further investigation of the platform’s dynamic behavior, particularly during operation in orchards and agricultural fields on sloping terrain, which is common in Armenia. Particular attention was given to changes in platform inclination during turning maneuvers on slopes and to the yaw moment generated about the vertical axis in four-wheeled robotic platforms with independently controlled wheels. Based on a mathematical model of the chassis incorporating elastic elements (springs) and damping elements (shock absorbers), the differential equation governing the motion of the platform’s center of mass was formulated, together with moment equations about the transverse and longitudinal axes passing through the center of mass. Analysis of these equations yielded the differential equations governing the platform’s vertical oscillations. Expressions were subsequently derived for the displacement, velocity, acceleration, and jerk associated with the platform’s free oscillations. These equations can be used to evaluate ride smoothness and determine the spring stiffness and shock-absorber damping coefficient required to achieve the desired suspension performance.
