Research in Agricultural Engineering - In Press
Investigation of Drying Kinetics of Residual Nuts Mixture Using Mathematical Models and Artificial Neural NetworksOriginal Paper
Tran Huu Duy, Nguyen Doan Kim Dang, Dang Nguyen Gia Han, Pham Tran Thanh Vy, Tran Ngoc Giau, Hong Van Hao, Nguyen Minh Thuy, Vo Quang Minh, Ngo Van Tai
The objective of this study was to evaluate the potential for sustainable reuse and the value of residual nuts mixture (RNM) byproducts (cashew nut, peanut, and soybean) after extraction. To investigate the drying kinetics, the RNM was dried at various temperatures (50 to 80 °C). The Balbay and ªahin model, which had a high coefficient of determination (R2) of 99.62–99.96%, a low root mean square error of 0.007–0.021, and χ2 of 0.001–0.005, was the one that best fit the experimental data out of the eight mathematical models that were used. Artificial neural networks showed higher and faster prediction capacity than mathematical models. The effective moisture diffusion coefficient (Deff) increased gradually with temperature, with an activation energy (Ea) of 13.67 kJ×mol-1. The RNM powder produced by the optimal drying process (60 °C for 3.75 h) has a bright color, high polyphenol content (2.68 mgGAE×g-1) and antioxidant activity, low moisture content (4.9%) and relatively high nutritional value, especially protein (27.27%), lipid (40.19%), and fiber (4.2%). Under these conditions, not only is efficient drying and preservation achieved, but the quality of the by-product powder is also maintained.
Quality Classification of Mangosteen using Image Processing and Machine LearningOriginal Paper
Usman Ahmad, Mardison Suhil
Mangosteen is one of important fruits for Indonesia with a good economic value both in domestic and international markets. Mangosteen has a soft and watery white flesh, consumed when the fruit is fully ripe, fresh with dark red color of peel and sometimes almost purple. For export quality grade, beside its tandard size and ripening stage, additional citeria are required by market such as the calyx with normal form and green color are necessary. Normal form and green color of calyx has no direct connection to internal quality of mangosteen, however it important for international market because they reflect the fruit freshness and esthetic value. The objective of this research is to develop quality classification of mangosteen into three classes; export, supermarket and others markets. Random Forest and Extreme Gradient Boosting models of machine learning were trained and validated using 300 samples using 26 image features, and then repeated using only 10 and 5 selected features for more robust classifier. The results of classification using all features were 90.0% and 86.7%, 86.7% and 85.0% using 10 selected features, and 86.7% and 75.0% using 5 selected features for Random Forest and Extreme Gradient Boosting respectively.
Harvester Service Life Impact on Sugarcane Field Losses and Product ContaminationOriginal Paper
Kanya Kosum
Mechanical sugarcane harvesting generates substantial material losses that are associated with equipment age. This study evaluated the relationship between harvester service life and operational efficiency by analysing field losses and product contamination across machines with varying operational histories (1, 14, 16, and 17 years) in Chaiyaphum Province, Thailand, using a randomised complete block design. Results indicate that 17-year-old machines exhibited 54% higher total losses (241.93 kg·ha⁻¹) compared to newer equipment (156.90 kg·ha⁻¹). Field losses were attributed primarily to base cutting operations (36%) and roller mechanisms (34%), collectively accounting for 70% of total losses. Contamination analysis revealed sugarcane tops as the predominant impurity source (57%Revenue loss analysis indicates excessive field losses from ageing equipment reduce farm profitability by 12–18%. The non-linear relationship between equipment age and performancedemonstrates that maintenance practices significantly influence degradation patterns, providing critical insights for optimising mechanical harvesting systems.
The Auto Sprinkler Rover: An Innovative Fertilizer Applicator for Sustainable Agriculture Original Paper
Wan khaima azira Wan mat khalir, ISMAIL AZIMAH, Nur Maizatul Akma Abdul Aziz (email: nurmaizatulakma011@ gmail.com)
Fertilizers are essential to agriculture since they provide crucial nutrients that stimulate crop growth, increase yield and improve yield quality. However, this practice requires the use of manual labour, which consumes energy and time as well as limits the effectiveness of fertilization methods and is environmentally harmful. These factors have led to the development of an innovative product based on the Internet of Things (IoT) called the Auto Sprinkler Rover, which is a remote controlled machine that operates automatically. It contains a 12 L water barrel that serves as a storage tank for liquid fertilizer. This study had designed a liquid sensor and light emitting diode to alert the user to the amount of liquid in the tank. The water pump facilitates the smooth spraying of liquid fertilizer from the pipe’s nozzle. Auto Sprinkler Rover differs from conventional methods because it integrates real-time monitoring via Wi-Fi, an automatic liquid level sensor as well as an automatic spraying system and ergonomic design that is manually operated. The Auto Sprinkler Rover was designed to be a cutting-edge technology that benefits sustainable crop cultivation for households, communities and the global agriculture industry.
Effect of maturity and colour parameters on cocoa opening force using the existing mechanical systemOriginal Paper
Amuaku Randy, Eric Amoah Asante, Emmanuel Y.H Bobobee, Jerry Oppong Adutwum, Godwin Amanor
The study assessed the mechanical and colourimetric properties of cocoa (Theobroma cacao L.) pods at various maturity stages to improve mechanised pod-opening efficiency. Thirty pods, each replicated three times, underwent compressive loading and colour analysis using the CIE Lab* colour space to correlate colour attributes with mechanical resistance. Pod maturity significantly (p < 0.05) influenced the opening force and compressive strength: unripe pods required the highest mean force (1,222 N) and strength (0.316 N.mm-2), while aged pods needed only 346 N and 0.094 N.mm-2. Transverse orientation yielded higher and more consistent force responses than longitudinal orientation, with ANOVA confirming significant differences (p < 0.05). Colour parameters, especially b* (yellowness), chroma (C), and hue angle (H°), were strongly and negatively correlated with mechanical properties (r ≥ –0.99, p ≤ 0.05), making them reliable non-destructive indicators of maturity. Partial Least Squares Regression (PLSR) models validated the predictive power of combined colour and mechanical data, with longitudinal orientation producing the highest model accuracy across all maturity levels. Combining colour-based maturity assessment with mechanical testing provides a robust framework for designing automated, maturity-sensitive cocoa pod-opening systems that optimise efficiency and reduce bean damage during postharvest handling. The research approach provided an excellent outstanding quantitative assessment.
Prediction of Postharvest Fruit Quality Using Machine Learning and Spectral Imaging DataOriginal Paper
Mardiantono Mardiantono, Zulfahrizal Zulfahrizal
Postharvest quality assessment of tropical fruits is crucial for preserving market value and ensuring consumer acceptance; however, conventional evaluation methods are destructive and unsuitable for real-time applications. This study developed and compared four machine learning (ML) algorithms—Support Vector Machine (SVM), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Artificial Neural Network (ANN)—for non-destructive prediction of fruit quality using spectral imaging data. Publicly available datasets of mango, papaya, and banana covering the 400–1000 nm wavelength range were employed. Both continuous attributes (soluble solids content, firmness, and moisture content) and categorical targets (ripeness and firmness classes) were evaluated. Spectral data were pre-processed using noise reduction, standardization, and dimensionality reduction techniques, and model performance was assessed using five-fold cross-validation. The results showed that ANN and XGBoost consistently outperformed SVM and RF for both regression and classification tasks. Feature importance analysis revealed that wavelengths in the 700–900 nm range were particularly informative, reflecting their association with sugar and water absorption characteristics in fruit tissues. These findings demonstrate the strong potential of integrating spectral imaging and machine learning for accurate, non-destructive fruit quality assessment. Nevertheless, further validation under field conditions and expansion to a wider range of fruit types are recommended to enhance model robustness and practical applicability.
Estimating changes in the Khisar glacier, using remote sensing data and GIS technologies for assesment of water use in agriculture (Surkhandarya valley, Uzbekistan)Original Paper
Shokhjakhon Khamidullaev, Rustam Oymatov, Ilhom Abdurahmanov, Ilkhom Aslanov
Climate change is speeding up the melting and retreat of glaciers, which is a big threat to water security in dry and semi-dry areas like Uzbekistan. To understand how glaciers affect regional hydrological systems and to come up with adaptive water management strategies, it is important to keep an eye on how they change over time. This study examines the temporal changes of the Khisar Glacier in the Surkhandarya Basin by combining remote sensing data from different times with GIS-based spatial analysis. We looked at Landsat images from 1990, 2000, 2010, and 2024 to see how the size of glaciers has changed and how that relates to weather and water variables. The results show that the glacier area has shrunk significantly, from 8.6 km² in 1990 to 5.1 km² in 2024, a 40.7% decrease over the past three decades. The mean annual temperature in the basin rose by about 1.9 °C during the same time, and the Surkhandarya River's average summer discharge fell by about 22%. These results show how closely rising temperatures, melting glaciers, and lower river flow are linked. They also show how vulnerable glacier-fed water systems are to climate change. Combining satellite observations with climate and hydrology data is a good way to keep an eye on glaciers and assess water resources over time. The GIS-based monitoring framework created in this study provides useful tools for planning how to adapt to climate change and manage water resources in a way that is good for the environment in the Surkhandarya region and other glacier-dependent basins in Central Asia.
Vetiver Grass for Road Slope Protection through Soil Erosion and Runoff Reduction in a Haor Area of Sunamganj DistrictOriginal Paper
Tahmina Akter, Saumik Acharya, Sayed Jaber Al-Hossain, Golam Kibria Robbani
Erosion control on slopes and embankments is a global concern, with vetiver planting emerging as an effective solution. This study investigated vetiver grass as a potential alternative for reducing runoff and soil erosion in protecting road embankments. In a small-scale model on a 1V:2H slope, simulated rainfall was applied through a perforated sheet. Runoff was collected, and eroded soil samples were dried using the sand bath method. Runoff and soil loss were compared with and without vetiver grass. Results showed that vetiver grass significantly reduced soil erosion, preventing about 87% more soil loss than bare soil. During 60 minutes of simulated rainfall, the bare soil plot lost per unit area is 27.196 kg.m-2, whereas the vetiver-covered plot lost only 3.606 kg.m-2. This soil retention is due to vetiver’s strong, deep root system and hedge-like structure that effectively holds the soil. Additionally, the runoff was reduced by 14% in plots covered with vetiver. For the 60 minutes of cumulative rainfall, the total runoff in the bare soil micro-plot was 81.45 cm, compared to 70.35 cm in the vetiver micro-plot. Therefore, vetivers are expected to be a sustainable and eco-friendly solution for runoff and erosion control within the lowland haor context.
Research on the Optimal Design and Process Parameters of a Castor Seed Cleaning MachineOriginal Paper
Elchyn Aliiev, Valentyn Holovchenko, Olha Aliieva
In the context of the modern EU bioeconomy, the use of industrial crops, particularly castor bean (Ricinus communis L.), is relevant for the production of industrial and energy products without competing with food crops. Castor oil is used for the production of biodiesel, lubricants, paints and coatings, cosmetics, and pharmaceuticals. In Ukraine, castor can be cultivated on low-yield soils, with seed yields ranging from 1.5 to 2.1 t/ha depending on the variety. However, the morphological features of the fruits and uneven ripening complicate the mechanization of harvesting and seed cleaning processes. Existing equipment designed for other crops cannot be applied due to the risk of damaging castor seeds. A design of a castor seed cleaning machine has been proposed, combining mechanical fruit shelling with aerodynamic separation of the mixture. The machine is equipped with an eccentric crushing cone, rubber linings, a pneumatic separation channel, a cyclone, and an automated control unit. As a result of numerical modeling and experimental studies of the process of separating and cleaning castor seeds, dependencies were obtained for the productivity of the developed machine Q, power consumption P, specific energy consumption E, fraction of unshelled fruits (segments) ξf, content of clean seed in the seed collector ψs-s depending on the distance between the reverse and crushing cones δ, rotation frequency of the crushing cone n, diameter of the feed opening Din, inclination angle of the crushing cone axis γ, and airflow velocity V. A multi-criteria optimization method was applied to find the optimal operating modes: δ = 10.8 mm; n = 282 rpm; V = 3.6 m/s; Din = 98 mm; γ = 3.6°; β = 20.3°. The following results were achieved: E = 0.0394 MJ/kg; Q = 163.4 kg/h; P = 1861 W; ξf = 0.099; ψs-s = 0.958. The obtained results confirm the efficiency of the proposed design for industrial implementation.
The effects of temperature on biogas production rate and purityOriginal Paper
Roger Jay Lamadrid De Vela, Romeica Noe Rimorin, Christian Mark Felix
This study investigated the effects of temperature on the performance of anaerobic digesters for biogas production. Digesters were filled with a 1:1 ratio of substrate to water, containing 15 kg of cow dung and 3 kg of crop waste, and maintained at temperatures of 50±2°C and 30±2°C, corresponding to the thermophilic and mesophilic biodigesters, respectively. The experiments run for 75 days, and biogas production rate and purity were measured. The thermophilic digester produced 48.4% more biogas and had a slightly higher pH (7.65) than did the mesophilic digester (7.37) by the end of the observation period. However, gas chromatography revealed that the CH4 and CO2 contents did not significantly differ between the two treatments. The CH4 concentration in the mesophilic environment was 42±10%, whereas that in the thermophilic environment was 53.5±10%. The CO2 composition was 32.5±1% and 35.5±1% for the mesophilic and thermophilic setups, respectively. These were supported by the wavelength (460 to 620 nm) of the flame color, indicating that the biogas from both set-ups is predominantly composed of methane. In conclusion, thermophilic anaerobic digesters may have a relatively high biogas production rate, but the biogas purity is not significantly different from that of mesophilic digesters.
Comparative study of Portable Vis-NIR Spectrometers for Corn Moisture Content Prediction Using Machine LearningOriginal Paper
Harki Himawan, Muhammad Dzakky Alghifari, Rut Juniar Nainggolan, Mochamad Bagus Hermanto, Nazmi Mat Nawi, Ken Abamba Omwange, Dimas Firmanda Al Riza
Non-destructive estimation of corn kernel moisture content is essential for determining the optimal harvest period. Although various spectrometer sensors are currently available, their predictive performance differs due to variations in spectral resolution and wavelength coverage. This study compared the performance of several portable spectrometer sensors with different wavelength ranges for predicting corn moisture content. Spectral data and reference moisture content were used to develop prediction models using Partial Least Squares Regression (PLSR) and Artificial Neural Network (ANN). Based on PLSR modelling, the AS7265X and C12880MA sensors produced the best performance, with coefficients of determination (R²) for training and testing reaching up to 0.90. Furthermore, ANN modelling yielded improved predictive accuracy, with the highest R² value of 0.95 obtained using the same sensor combination. These results demonstrate that portable spectrometers show strong potential for non-destructive field-based prediction of corn moisture content and can serve as a reliable indicator for determining optimal harvest timing.