Predicting the methane yield of lignocellulosic biomass in mesophilic solid-state anaerobic digestion based on feedstock characteristics and process parameters

Abstract

In this study, multiple linear regression (MLR) and artificial neural network (ANN) models were explored and validated to predict the methane yield of lignocellulosic biomass in mesophilic solid-state anaerobic digestion (SS-AD) based on the feedstock characteristics and process parameters. Out of the eleven factors analyzed in this study, the inoculation size (F/E ratio), and the contents of lignin, cellulose, and extractives in the feedstock were found to be essential in accurately determining the 30-day cumulative methane yield. The interaction between F/E ratio and lignin content was also found to be significant. MLR and ANN models were calibrated and validated with different sets of data from literature, and both methods were able to satisfactorily predict methane yields of SS-AD, with the lowest standard error for prediction obtained by an ANN model. The models developed in this study can provide guidance for future feedstock evaluation and process optimization in SS-AD.