In the BeProMod project, a novel framework for modeling multi-scale biorefinery processes was implemented. The established method of dynamic flux balance analysis (DFBA) was extended to cover non-linear cellular objective functions, and an extreme-ray based reformulation is proposed to guarantee certain regularity conditions in the Karush-Kuhn-Tucker formalism which is applied for solving the resulting constrained differential-algebraic equations. In a case study, the metabolic network of Corynebacterium glutamicum was extended to include amino acid and nucleotide synthesis pathways, a D-xylose assimilation pathway and a pathway for succinate production. This metabolic network was then coupled with a dynamic extracellular process model for the fermenter, which includes detailed calculation of the pH value in the medium. The resulting DFBA model is converted to the MODELICA language and can be directly solved using specifically tailored numerical solvers. An integrated hierarchical multi-scale model of a biorefinery based on the NGP reference process was established by combining models of the Organocat pretreatment process, fermenter, and downstream processing units.
As the final biorefinery model contains a high number of possible (bio-) chemical process steps, particular emphasis was placed on ease of use and on reusability of the building blocks. The most relevant unit operations were implemented including models for mixer, splitter, flash, decanter and distillation column. The developed models are in general able to describe both steady-state and dynamic processes and consider both mass and energy balances. This framework is not limited to the considered case study but can be used for modeling dynamic processes from other application fields. When solving the given problem instance numerically, further novel approaches were required and successfully applied in order to resolve issues arising from non-smoothness in the underlying mathematical models. In addition, the software infrastructure has been extended to exploit parallelism and adjoint mode algorithmic differentiation.
Participating Core Groups (CG)
Prof. Dr. Wolfgang Wiechert (with Dr. Eric von Lieres and Dr. Stephan Noack)
Forschungszentrum Jülich GmbH
Institute of Bio- and Geosciences
Phone: +49 2461 61 - 3118
Fax: +49 2461 61- 3870
Prof. Alexander Mitsos, Ph.D.
Fakultät 4, Aachener Verfahrenstechnik, Systemverfahrenstechnik– SVT RWTH Aachen
Prof. Dr. Uwe Naumann
Fakultät 1, Informatik, Software and Tools for Computational Engineering – STCE RWTH Aachen
01.07.2014 – 30.06.2017
The total budget of BEProMod is € 618.648 .00. BEProMod is part of the NRW-Strategieprojekt BioSC and thus funded by the Ministry of Innovation, Science and Research of the German State of North Rhine-Westphalia.
Lotz, J, Naumann, U, Hannemann-Taḿas, R, Ploch, T and Mitsos, A (2015). Higher-order discrete adjoint ode solver in c++ for dynamic optimization. Procedia Computer Science 51: 256-265.
Lotz, J, Schwalbach, M and Naumann, U (2016). A case study in adjoint sensitivity analysis of parameter calibration. Procedia Computer Science 80: 201-211.
Safiran, N, Lotz, J and Naumann, U (2016). Algorithmic differentiation of numerical methods: Second-order adjoint solvers for parameterized systems of nonlinear equations. Procedia Computer Science 80: 2231-2235.
Zhao, X, Noack, S, Wiechert, W and Lieres, EV (2017). Dynamic flux balance analysis with nonlinear objective function. J Math Biol