Prediction of Sugar Yields During Hydrolysis of Lignocellulosic Biomass Using Artificial Neural Network Modeling

Abstract

The present investigation was carried out to study application of ANN as a tool for predicting sugar yields of pretreated biomass during hydrolysis process at various time intervals. Since it is known that biomass loading and particle size influences the rheology and mass transfer during hydrolysis process, these two parameters were chosen for investigating the efficiency of hydrolysis. Alkali pretreated rice straw was used as the model feedstock in this study and biomass loadings were varied from 10% to 18%. Substrate particle sizes used were <0.5 mm, 0.5–1 mm, >1 mm and mixed size. Effectiveness of hydrolysis was strongly influenced by biomass loadings, whereas particle size did not have any significant impact on sugar yield. Higher biomass loadings resulted in higher sugar concentration in the hydrolysates. Optimum hydrolysis conditions predicted were 10 FPU/g cellulase, 5 IU/g BGL, 7500 U/g xylanase, 18% biomass loading and mixed particle size with reaction time between 12–28 h.