(390a) Development of An Integrated Dynamic Model for Bioethanol Production From Lignocellulosic Biomass | AIChE

(390a) Development of An Integrated Dynamic Model for Bioethanol Production From Lignocellulosic Biomass


Gernaey, K. V. - Presenter, Technical University of Denmark
Sin, G. - Presenter, Technical University of Denmark
Morales-Rodriguez, R. - Presenter, Denmark Technical University
Meyer, A. S. - Presenter, Technical University of Denmark

Biofuel production from lignocellulosic biomass is a complex process. A typical flowsheet of bioethanol production from lignocellulosic biomass involves a number of unit operations such as physical/chemical pre-treatment, enzymatic hydrolysis, fermentation, separation among others. Improving both the design and the energy efficiency of bioethanol production processes are among the chief challenges of demonstrating the feasibility/viability of bioethanol as alternative renewable energy resource. Development and transfer of these processes from proof-of-concept to industrial scale are mainly done on an empirical basis, and typically rely on experiences from conventional one-pot conversion processes. This approach is rather inefficient and costly in terms of time and resource investments. This study introduces a model-based simulation framework for biofuel process design (see Figure 1). The hypothesis is that simulations with a dynamic mechanistic model will facilitate rational process development and boost innovative designs. Central to this design approach is the availability of reliable dynamic models of all the reactions and unit processes involved in the biofuel production process. Figure 1. A model-based simulation framework for design of new enzyme processes. The proposed model-based design framework is iterative and realized in two complementary phases (see Figure 1). In the first phase, one aims at developing a reliable model for describing the rates of each process units. Once a reliable model for each unit is obtained, the next phase is to develop an integrated model of the entire bio-ethanol production process by connecting/bringing together each process unit in one modelling platform. Such integrated process models form the basis to help generate and evaluate innovative process flowsheets. A couple of integrated models were developed for bioethanol production systems1,2. These models employed short-cut models of energy and mass balances to mathematically describe each process unit and most importantly were steady-state. Such steady-state integrated models were indeed valuable for evaluating alternative process flowsheets in view of energy efficiency (and in some cases formed the basis for optimization of the design by nonlinear programming). However there are limitations of using steady-state models, which are the following: ? One cannot obviously evaluate the dynamic interactions between process units notably those between the hydrolysis and fermentation units. The degree of interaction will depend on the operational mode (batch versus continuous) and also on the reactor configuration (whether it is a combined or separate units of hydrolysis and fermentation). Furthermore, bearing in mind that fermentative organism (e.g. Saccharomyces cerevisiae) are inhibited by elevated product levels in the fermentation reactor (e.g. ethanol) as well as some intermediate products formed during the currently used pre-treatment methods, the design (the calculated yield and energy efficiency target among others) based on steady-state models will be at best too optimistic rather than realistic. ? Steady-state models cannot be used for up or down-scaling purposes as they lack the necessary kinetics needed to calculate the time-constants of different processes. ? One cannot test and evaluate feasibility of control strategies needed to operate the systems at a preconceived operational trajectory. In short, while the design of bioethanol processes using lignocellulosic biomass based on steady-state models is definitely valuable, it needs however to be verified and further refined by dynamic simulations using integrated dynamic models. This is precisely the objective of this study to develop an integrated dynamic model for a bioethanol production process using lignocellulosic biomass. To this end, the dynamic model for each process unit including pre-treatment (in-house developed model), hydrolysis (a modified and validated NREL model3), fermentation (a validated in-house model4) and separation (distillation column) are brought together in one platform (Matlab Simulink) and connected using appropriate model interfaces where necessary. The integrated dynamic model was then used to evaluate and verify the steady-state ethanol yield and energy efficiency of the process. Furthermore, the model simulations will be used to investigate the interaction between the hydrolysis units and fermentation units such that an appropriate controller can be developed to guarantee optimal operation of these units. Finally, it is likely that a carefully developed model-based simulation framework will identify new avenues for improved development of biofuels processing strategies using lignocellulose. References 1. CA Cardona Alzatea, OJ Sanchez Toroa. Energy consumption analysis of integrated flowsheets for production of fuel ethanol from lignocellulosic biomass. Energy 2006; 31:2447?2459. 2. Karuppiah R, Peschel A, Martin M, Grossmann IE, Martinson W, Zullo L, Energy (2008) Optimization for the Design of Corn-Based Ethanol Plants. AIChE Journal 2008; 54:1499-1523. 3. Sin G ,Meyer AS, Gernaey KV (2009) Are mechanistic cellulose-hydrolysis models reliable for use in biofuel process design? ?uncertainty and sensitivity analysis.). In: Design for Energy and the Environment. Proceedings of the 7th international conference on the foundations of computer-aided process design (FOCAPD). 4. Sin G, Ödman P, Petersen N, Lantz AE, Gernaey KV (2008) Matrix notation for efficient development of first-principles models within PAT applications: Integrated modeling of antibiotic production with Streptomyces coelicolor. Biotechnol. Bioeng. 2008; 101: 153?171.