The prediction and optimization of biomass compatibility ratios by machine learning for enhancing methane production...

This study presents the development of an ACoD model using machine learning techniques to predict the optimal input parameters for maximizing methane production. The results indicated that the machine learning model achieved a high predictive accuracy, with maximum average R2 and RMSE values of 0.951 and 0.57, respectively. The key factors influencing methane production were carbon content, reaction time, and feedstock ratio, which contributed over 90% to the output variability. Optimization objectives based on cumulative and daily methane production yielded optimal feedstock ratios of 2:5 and 1:9, respectively. Experimental validation of these predictions demonstrated strong alignment with the model. These findings provide valuable insights into constructing efficient ACoD models and optimizing feedstock ratios, offering a foundation for improving methane production efficiency in practical applications.