Application of Bayesian Modelling and Central Composite Design on Sustainable Biogas Production through Anaerobic Co-digestion of Tannery Residues
摘要
Tannery solid waste management posed significant environmental challenges, particularly due to the high organic load and potential toxicity of untreated fleshings. This study explored the Anaerobic co-digestion (AcoD) of chromium-free tannery fleshing, cow dung, and sewage to optimize biogas production under varying operational conditions. A multi-phase experiment was adopted, initially comparing room and controlled temperature conditions, followed by evaluations across mesophilic to lower thermophilic ranges (30–45 °C) and nine different substrate mix ratios. The highest methane yield and digestion performance were achieved at 45 °C with a substrate including a mix of fleshing, sewage, and cow dung as 1:2:1. Process parameters, primarily pH, COD reduction, volatile solid (VS) reduction, and volatile fatty acids-to-alkalinity ratio, were monitored to assess system stability. A Bayesian Network was constructed to model interdependencies among these variables, identifying temperature, initial volatile solids, and substrate-to-inoculum (S/I) ratio as key drivers of gas production. A sensitivity analysis based on mutual information and belief variance further highlighted their influence, providing a probabilistic framework for process optimization with strong interdependence, such as S/I ratio-pH (75.0%), S/I ratio-initial VS (61.5%), and VFA/Alkalinity ratio-temperature (61.6%). Thereafter, a Central Composite Design coupled with desirability analysis was used to identify optimal operational settings, followed by validation experiments, resulting only 5.6 ± 0.5% error. This study demonstrated a successful integration of experimental results and probabilistic modeling to enhance energy recovery from hazardous tannery waste, and its findings contributed to the sustainable waste-to-energy strategy as well as the circular bioeconomy.