Multi-scale uncertainty analysis – A tool to systematically consider variability in lignocellulosic bioethanol processes

Other conference contribution, 2018

Bioethanol production processes from lignocellulosic raw materials are highly prone to batch-to-batch variations. For example, raw material compositions and enzymatic activities required to release fermentable sugars from lignocellulose vary significantly between batches. To develop lignocellulosic biofuel processes and evaluate their performance regarding economics and sustainability consistently, tools are required to cope with this variability.

 

In this presentation we will propose a multi-scale uncertainty analysis strategy to propagate input variability throughout system scales. In a first step, we use meta-data obtained from literature to define uncertainties in the process inputs. Utilizing these meta-data, uncertainty analysis is performed on a macro-kinetic model by sampling from the defined uncertain input space. The results of this uncertainty analysis are transferred to process simulations to analyze the impact of input uncertainties on the process mass- and energy balances, and on the economics of building this type of bioprocess. The generated data from process simulations (mass flows, energy integration, and economic data) are in the next step extracted and used as input to an environmental impact assessment of the process. This is done whilst keeping the simulation and systems modeling parameters constant, thus the input variability is propagated throughout the different system scales. The data generated in this integrated approach will then be compared with the variations and uncertainties observed with relevance to the estimated parameters in the process simulation and environmental impact assessment. Based on this consistent strategy, we can analyze the impact of input variability from different system perspectives, identify important bottlenecks for development, and suggest robust and sustainable process designs for different conditions and under given uncertainties.

 

In a case study we demonstrate how integrated kinetic modeling (in Matlab), process simulation (in SuperPro Designer), and environmental impact assessment together with statistical analysis can be used for assessing how variability in enzymatic activities in bioethanol production can be propagated throughout system scales. A macro-kinetic model is used to describe the enzymatic breakdown of lignocellulose-derived polysaccharides into fermentable sugars (saccharification) and the simultaneous fermentation to bioethanol. We discuss the integration of the simulation results of the macro-kinetic model into the flowsheeting software for mass and energy balance generation, and then further on to assess environmental impacts of the process. We will evaluate different process designs regarding their robustness towards input variability. Finally, we also show how propagated uncertainties at different system scales can be integrated to design experiments at laboratory scale so that these focus on the most important parameters for developing robust kinetic models, and include the parameters that are most important for sustainable design of processes and value chains.