Multiscale modeling and predictive control of cellulose accessibility in alkaline pretreatment for enhanced glucose yield

Abstract

Even though overcoming biomass recalcitrance through chemical pretreatment significantly improves the glucose yield during biorefinery process, in-depth understanding of the pretreatment has been impeded by the lack of robust kinetic models that capture the microscopic properties of biomass (e.g., cellulose accessibility). In this work, a multiscale model is developed to describe the dynamic evolution of cellulose accessibility during pretreatment. Then, a multi-reaction kinetic model is adopted from the literature and modified to describe the glucose yield of biomass from enzymatic hydrolysis. Then, model approximation techniques are employed to the pretreatment and hydrolysis models to ease the inherent model complexity. Utilizing the approximate models, a model-based feedback controller is designed to enhance the cellulose accessibility while minimizing the heat during alkaline pretreatment. The implementation of the control framework improved the glucose yield by 46% while consuming 7.2% more heat energy compared to a conventional constant-temperature pretreatment method.