The aim of this study is to identify an alternative approach for simulating the in vitro fermentation and quantifying the production of rumen methane and rumen acetic acid during the rumen fermentation process with different total mixed rations. In this experiment, dietary nutrient compositions (neutral detergent fiber (NDF), acid detergent fiber (ADF), crude protein (CP), and dry matter (DM)) were selected as input parameters to establish three prediction models for rumen fermentation parameters (methane and acetic acid): an artificial neural network model, a genetic algorithm-bp model, and a support vector machine model. The research findings show that the three models had similar simulation results that aligned with the measured data trends (R² ≥ 0.83). Additionally, the root mean square errors (RMSEs) were ≤1.85 mL/g in the rumen methane model and ≤2.248 mmol/L in the rumen acetic acid model. Finally, this study also demonstrates the models’ capacity for generalization through an independent verification experiment, as they effectively predicted outcomes even when significant trial factors were manipulated. These results suggest that machine learning-based in vitro rumen models can serve as a valuable tool for quantifying rumen fermentation parameters, guiding the optimization of dietary structures for dairy cows, rapidly screening methane-reducing feed options, and enhancing feeding efficiency.

The temperature at the base of the ear is highly correlated with the core body temperature of sheep and responds sensitively to febrile conditions, making it a valuable indicator of sheep health. In northern China, the closed housing environment during winter increases the incidence of seasonal diseases such as upper respiratory infections and pneumonia, which severely affect the economic efficiency of sheep farming. To address this issue, this study proposes an early-warning method for winter diseases in sheep based on ear-base temperature. Ear temperature, body weight, and environmental data were collected, and Random Forest was employed for feature selection. Bayesian optimization was used to fine-tune the hyperparameters of a one-dimensional convolutional neural network to construct a predictive model of ear-base temperature using data from healthy sheep. Based on the predicted normal range, an early-warning strategy was established to detect abnormal temperature patterns associated with disease onset. Experimental results demonstrated that the proposed method achieved a high detection rate for common winter diseases while maintaining a low false positive rate, and validation experiments confirmed its effectiveness under practical farming conditions. Combined with low-cost temperature-sensing ear tags, the proposed approach enables real-time health monitoring and provides timely early warnings for winter diseases in large-scale sheep farming, thereby improving management efficiency and economic performance.

Accurate state of charge (SOC) estimation is essential for the reliable operation and energy management of electric agricultural machinery, particularly electric tractors operating under complex field conditions. This study aims to improve SOC estimation accuracy and robustness by proposing a hybrid method that integrates a gated recurrent unit (GRU) neural network with a fractional-order extended Kalman filter (FOEKF). The GRU model is employed to capture the nonlinear behavior of lithium-ion batteries, while the FOEKF is used to mitigate noise and compensate for model uncertainties, forming a coupled data-driven and model-based framework. Experiments were conducted on lithium-ion batteries for electric tractors under hybrid pulse power characterization (HPPC) conditions at 15 °C, 25 °C, and 35 °C. These experiments can effectively simulate the dynamic power fluctuation characteristics of the battery caused by variations in electric tractor operating conditions during agricultural operations in different seasons. Experimental results demonstrate that, compared with conventional GRU and FOEKF methods, the proposed GRU-FOEKF method achieves lower estimation errors and improved robustness. In particular, at 25 °C, the proposed method achieves an MAE of 0.9% and an RMSE of 1.1%, outperforming the compared algorithms. These findings indicate that the proposed strategy is a feasible and effective solution for battery management systems in electric agricultural machinery, contributing to the development of smart and sustainable agriculture.