Fuzzy-enhanced modeling of lignocellulosic biomass enzymatic saccharification

Abstract

The enzymatic hydrolysis of lignocellulosic biomass incorporates many physico-chemical phenomena, in a heterogeneous and complex media. In order to make the modeling task feasible, many simplifications must be assumed. Hence, different simplified models, such as Michaelis-Menten and Langmuir-based ones, have been used to describe batch processes. However, these simple models have difficulties in predicting fed-batch operations with different feeding policies. To overcome this problem and avoid an increase in the complexity of the model by incorporating other phenomenological terms, a Takagi-Sugeno Fuzzy approach has been proposed, which manages a consortium of different simple models for this process. Pretreated sugar cane bagasse was used as biomass in this case study. The fuzzy rule combines two Michaelis-Menten-based models, each responsible for describing the reaction path for a distinct range of solids concentrations in the reactor. The fuzzy model improved fitting and increased prediction in a validation data set.