New process configurations for lignocellulosic ethanol production: economic optimization and uncertainty analysis.

In lignocellulosic ethanol production processes, any change made in the saccharification and fermentation area impacts capital and operating costs of downstream areas. Despite several plant–scale techno-economic studies of the most common process configurations have been reported, the number of process configurations assessed by these reports is rather limited. Moreover, the uncertainty in technical (yields and productivities) and economic (investment and operating costs) has been largely neglected.

In this work we address a systematic, optimization-driven, search of economically improved process configurations in the saccharification and fermentation section of a stand-alone lignocellulosic ethanol plant using corn stover as feedstock. Uncertainty in the experimental data and in kinetic models, is represented as confidence intervals of the parameters of kinetic models for the saccharification and fermentation stage. Each possible process configuration is integrated into a plant scale process simulation model including distillation, waste water treatment and heat and power production. Moreover, an uncertainty analysis is performed over a subset of optimized process configurations to identify those configurations that are economically near-optimal and that can withstand uncertainty in model parameters. Processes were simulated in Aspen HYSYS^TM while process optimization and uncertainty analysis were performed in a tailored optimization system integrating MATLAB^® and GAMS, able to solve disjunctive programming optimization problems containing tens of thousands variables.

A number of new process configurations for lignocellulosic ethanol production from corn stover were found, that are economically near-optimal and that show a reduced sensitivity of economic indicators towards uncertainty in model parameters of the saccharification and fermentation stages.