Xiaoyu Tang

Aerobic composting shows state-dependent dynamics in key parameters such as temperature, moisture content, oxygen concentration, and pH, and these variables are strongly coupled over time. This coupling makes accurate state identification and process regulation challenging when relying on single-parameter thresholds or experience-based control. To enable robust recognition of composting states throughout the process, we propose an IRRTO-optimized CNN–LSTM–attention model (IRRTO–CNN–LSTM–attention). The model uses a convolutional neural network (CNN) to extract discriminative multivariate features, a long short-term memory (LSTM) network to model temporal dependencies, and an attention module to adaptively emphasize informative features. To address the hyperparameter selection challenge, the Rapidly-exploring Random Tree Optimizer (RRTO) was introduced and further enhanced via four strategies (fluctuating attenuation adaptive regulation, dual-mode guided update, dynamic dimension adaptive perturbation, and dual-mechanism adaptive perturbation regulation), forming the improved IRRTO. The proposed approach was validated using sensor data from windrow composting of pig manure and corn straw. The IRRTO–CNN–LSTM–attention model achieved an overall accuracy of 98.31% in classifying the four states (mesophilic/heating, thermophilic, cooling, and abnormal) on the independent test set, which was 3.39 percentage points higher than the RRTO-based model. These results suggest that the proposed method can accurately identify composting states and support early warning and state-specific regulation in practical aerobic composting systems.

Moisture content varies continuously during aerobic composting, which changes material flowability and can limit the use of a single set of discrete element method (DEM) parameters. To address this issue for a multi-component corn straw–pig manure mixture, we developed a rapid calibration workflow covering a moisture content range of 29–80%. Angle of repose (AoR) images were obtained using a cylinder-lifting test. To improve robustness for irregular pile contours, we proposed an AoR extraction method that combines LOESS smoothing with least-squares line fitting. Key DEM contact parameters affecting AoR were screened using a Plackett–Burman design, and their effective ranges were refined using a steepest-ascent test. A Box–Behnken design was then used to establish a response surface linking AoR to the significant DEM parameters. In addition, a polynomial relationship between moisture content and AoR was fitted and coupled with the AoR-parameter response surface to predict key DEM parameters directly from moisture content. Validation results showed that the predicted AoR exhibited a relative error below 10% across the tested moisture contents. An independent baffle-lifting validation test yielded a relative error below 5%. Overall, this workflow provided a practical strategy for setting DEM simulations of composting feedstocks under variable moisture content and supports numerical analysis and structural optimization of composting-related machinery.

About 30–40% of lignin-rich unhydrolyzed solids (UHS) are left behind after subjecting lignocellulosic biomass to thermochemical pretreated processes followed by enzymatic hydrolysis (EH) to produce sugars that are fermented to fuels and chemicals in a biorefinery. Ammonia Fiber Expansion (AFEX) is one of the leading alkaline pretreatment processes that use volatile ammonia that can be recovered and reused beneficially for the environment. In this work, we used AFEX-EH-UHS which are produced after subjecting corn stover to AFEX followed by EH and contain carbohydrates, ashes, and other impurities that are detrimental to the conversion of lignin to high-value products. In the study, we discovered that ~80% of the carbohydrates present in AFEX-EH-UHS were hydrolyzed and consumed during the AD process. The resulting solids, hereafter called AD-UHS, were subjected to lignin extraction using different combinations of solvents under reflux conditions. The solvent-extracted lignin was subjected to thermogravimetry, nuclear magnetic resonance (NMR) spectroscopy, and molecular weight analysis. Among the solvents, acetic acid could produce 95% pure lignin with some chemical modification, while aqueous ethanol was able to produce 80% pure lignin without any chemical modification.