Categories
Nevin Manimala Statistics

Application of bokashi for enhancing anaerobic digestion and sustainable biogas production: recurrent neural network (RNN) modeling implementation

Environ Sci Pollut Res Int. 2025 Nov 28. doi: 10.1007/s11356-025-37176-8. Online ahead of print.

ABSTRACT

Anaerobic digestion is an effective technology for converting organic waste into biogas while reducing environmental pollution. This study investigates the impact of co-digesting waste-activated sludge (WAS) with wheat straw, rice straw, and bokashi on biogas production. Nine anaerobic batch reactors were operated under mesophilic conditions (35 °C), incorporating different proportions of bokashi (1% and 2%) along with rice and wheat straw (4%). The results revealed that reactors supplemented with wheat and rice straw exhibited higher biogas production than the control reactor (sludge only). Wheat straw outperformed rice straw in improving biogas yield, total solids (TS) reduction, total volatile solids (TVS) degradation, and chemical oxygen demand (COD) removal. The addition of bokashi enhanced biogas production, confirming its role in accelerating organic matter breakdown. The maximum biogas yield was observed in the reactor containing sludge co-digested with wheat straw and 2% bokashi, which generated three times more biogas than the control. This reactor also exhibited the highest degradation rates of TS (57.83%), TVS (66.37%), and COD (71.53%). Furthermore, pH remained stable within the optimal range across all reactors, ensuring a balanced digestion process. Statistical analysis revealed significant correlations between organic matter degradation (COD, TS, TVS reduction) and biogas production, demonstrating that effective substrate decomposition improves biogas yield. The recurrent neural network (RNN) model was applied to experimental data to predict biogas production. With an exceptionally low root mean square error (RMSE) of 0.0041, R2 close to 1, and MAE 0.0117, the model exhibited excellent accuracy and reliability in generating precise predictions.

PMID:41313517 | DOI:10.1007/s11356-025-37176-8

By Nevin Manimala

Portfolio Website for Nevin Manimala