Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

A model-based approach for multi-objective optimization of solid oxide fuel cell systems

: Reuber, S.; Strelow, O.; Dittmann, A.; Michaelis, A.

European Fuel Cell Forum -EFCF-:
10th European Solid Oxide Fuel Cell Forum, SOFC 2012. Proceedings : Fundamentals, Materials, Systems, Applications; 26 - 29 June 2012, Lucerne/Switzerland
Oberrohrdorf/Switzerland: European Fuel Cell Forum, 2012
European Solid Oxide Fuel Cell Forum (SOFC) <10, 2012, Luzern>
Conference Paper
Fraunhofer IKTS ()
Mehrzieloptimierung; electrical output; process layout; thermodynamische Leistung

Fuel cell system design is a challenging endeavour due to the many feasible process configurations, the high level of system integration and the resulting component interactions. Multiple economic and environmental design criteria, that often conflict each other, need to be observed simultaneously prior to extensive hardware testing. In such cases process simulations can aid significantly to study system effects while keeping development time short and costs low. In fuel cell literature optimization of cell design or operational parameters with respect to only objective is much more common than optimization of the process structure itself. Within this work an approach from process system engineering has been extended to allow for multi-objective optimization of fuel cell systems. Thus a comparison of different layouts is quickly possible. The method will be presented for a SOFC based power generator with electrical output of 5 kW el. The structure of the process layout is analyzed and transferred into a matrix equation of mass and energy balances equations. Free design variables are extracted by elementary matrix manipulations. Based on these variables a steady state process simulation is set up to describe the thermodynamic performance of the fuel cell system including thermal and fluidic interactions. The process model can be easily validated to experimental data. For economic evaluation the simulation roughly computes capital costs of key components. Pareto optimum for specific costs and net efficiency is numerically computed by a robust genetic algorithm from Matlab. It is shown that a small decline of 2% in efficiency leads to cost saving up to 15 %. With the approach an evaluation of prospective design concepts in terms of efficiency and capital costs is quickly feasible. A sensitivity analysis can assist target-orientated hardware development and focuses on critical system components.