General modeling of the enzymatic hydrolysis process of lignocellulosic materials using approaches based on Artificial Neural Networks

Abstract

The computational optimization of a process for obtaining 2G ethanol requires a model of the enzymatic hydrolysis stage that can be integrated into the rest of the process. Currently, there are many projects that represent its actions under certain experimental conditions. However, their predictions can be used as a guide and under operating conditions similar to those studied.The general models development can be extrapolated to different raw materials and different operating conditions is still a challenge, since the different processes involved in enzymatic hydrolysis are superficially known. However, for this reason, the experimental information available on the enzymatic hydrolysis of lignocellulosic materials can be used, in combination with non-conventional modeling methodologies in the field, as is the case of Artificial Neural Network modeling.In the present work, the performance of two approaches based on the Artificial Neural Networks model is analyzed to explain the behavior of the enzymatic hydrolysis process of different raw materials subjected to different pretreatments to obtain a general predictive model.