Modeling and Simulation
In modeling fuel cell stacks, strong non-linear coupled processes should be taken into account on different spatial and time scales. These different scales mean that it is not possible to use just one single modeling approach.
A suitable model and simulation tool must therefore be selected for each individual problem. On the one hand, we explore basic relationships of physics, chemistry, fluid dynamics and thermomechanics. On the other hand, we work towards advanced designs and processes for new cells stacks and reformers. To this end, intensive cooperation with IEK-3’s new Physicochemical Fuel Cell Laboratory is important for determining structure–activity relationships using both mathematical and experimental methods. In particular, the Jülich supercomputers ensure simulation results of excellent quality permitting predictions of the operating behavior of new detailed constructions for complex energy converters. The processes are simulated from the material scale (?m) to the cell scale (mm), the stack scale (cm) and the system scale (m).
Modeling on the System Level
Stationary process simulations of chemistry and thermodynamics are carried out using the Pro/II program for on-board power supply systems based on HT-PEFCs and autothermal reforming. In addition to core system components, the heat exchangers for integrated heat recovery as well as pumps and compressors for overcoming the system pressure loss are also taken into consideration. New process flows can thus be tested for plausibility and optimum process layouts can be identified. The evaluation is based, among other factors, on the total heat-exchanging surface, the water balance, the partial efficiencies and the overall system efficiency. Model-based optimization also includes the improvement of heat integration by means of the pinch-point method as well as sensitivity analyses with respect to critical process parameters.
Realistic transient operating conditions, for example during the start-up process or load changes, are also represented by means of system modeling. Dynamic system modeling is thus the basis for developing operating and control strategies. This also includes our own models developed for the MATLAB/Simulink simulation environment. Additionally required components (e.g. a starting device) can be identified by this method. Components can be assessed at different operating points, not only the one they are designed for.
Modeling on the Stack Level
The simulation of fuel cells and reformers is described as an example of modeling on the stack level.
Operating large fuel cell stacks requires active temperature management for optimum operating performance, since the heat resulting from the reaction needs to be removed. Measurements with temperature sensors integrated in a measuring plate deliver data on the local temperature distribution inside fuel cell stacks.
Alternatively, the temperature distribution across HT-PEFCs can also be studied using numerical methods. The aim is to optimize the properties of fuel cell stacks. This requires fluid dynamics and heat transfer to be combined with the influences of reaction kinetics. A suitable model is therefore created of the HT-PEFC to be investigated and then simulated with the CFD software ANSYS FLUENT.
Overall models of the reactors are created by CFD-assisted design of the core components for fuel processing. The purpose of modeling with ANSYS FLUENT is to identify the thermodynamic interactions between feedstock preparation, reaction and postprocessing within the reactors. The creation of overall models is therefore a basic precondition for designing integrated heat exchangers and optimizing them for a compact component design. Due to the different physical phenomena modeling of an entire core component requires a standard model library, an integration of turbulence models and models for multiphase flows. Such models include the description of heat transfer and phase transition phenomena as well as models for describing chemical reactions. These complex modeling tasks are performed on the JUROPA supercomputer at Jülich.
Modeling on the Cell Level
As in the case in stack modeling, on the cell scale the simulation of individual cells in a fuel cell stack is considered as well as the components of a reformer.
An HT-PEFC stack consists of several individual cells. On the level of these cells, the local distribution of gases inside the active area of the membrane electrode assembly plays a central role. The resulting current density distribution and temperature distribution are major criteria for assessing an optimum configuration.
Partial models of the different reactors are designed and calculated as the basis for CFD modeling of the core components for fuel processing. Creating partial models serves both to design components and to create and validate models. Individual physical phenomena are mapped and studied with a high spatial and temporal discretization. Partial CFD models are used, among other purposes, for studying chemical reactions, highly dynamic injection processes (see Figure 3) and phase transition phenomena at heat exchanger walls. The CFD simulations are validated using laboratory experiments and flow visualization experiments by means of high-speed microscopy. Calculating the highly discretized partial models requires the capacity of the JUROPA supercomputer at Jülich.
In gas diffusion layers (GDL), transport processes are simulated using the Lattice-Boltzmann method. The aim of this work is to identify the influences of the GDL structures on the one and two phase transport processes.
The image shows the simulation of steam transport through a section (0.7 mm x 0.7 mm) of the GDL of a high-temperature polymer electrolyte fuel cell (HT-PEFC). Supercomputers are a necessary prerequisite for carrying out this kind of simulation.