Fuel Cell System Simulation

News From AVL’s Virtual Fuel Cell Development

The automotive industry has been shaped by the controlled use of combusting fossil fuels and the conversion of heat into mechanical power since its very beginning. The next (r)evolutionary stage in propulsion, however, is coming from fuel cells. A fuel cell is an electrochemical cell that converts the chemical energy of the fuel and an oxidizer into electrical energy through redox reactions, producing only water as exhaust product.

Two things are crucial for this process. There’s the fuel stack itself, where the electrochemical and thermodynamic processes take place. Equally important is the Balance-of-Plant (BoP). This consists of the gas paths for fuel and air supply, a cooling circuit for thermal conditioning and an electric circuit that transports the electric power from the stack to the electrical drive unit and the battery. Various control functions, sensors and actuators optimize this operation.

From a modeling perspective, a fuel cell system is a textbook example of a multi-physical, multi-domain, multi-timescale system, fully covered by the modeling and simulation capabilities of our system simulation tool, AVL CRUISE™ M, which offers an extensive library of components for each of the vehicle and powertrain related domains. For Protone exchange membranes (PEM) fuel cell system simulation, AVL CRUISE M offers native, built-in models for the stack and all major BoP components, such as compressor, humidifier, water separator and injector/ejector. The BoP components are modeled based on a fully physical basis to account for their real transient behavior under dynamic operation.

For PEM fuel cell stack and system development, virtual integration and calibration, AVL CRUISE M offers a unique electrochemical reduced dimensionality (RD) model. It employs a quasi-dimensional gas flow model, which allows a 1D volume discretization at low computational demand. The RD electrochemical model solves the transport processes through the membrane electrode assembly in a quasi-1D manner, and it takes into account nitrogen crossover and transient water formation. This offers an optimum compromise between modeling depth and computational performance, compliant with real-time capabilities, making it predestined for system performance modeling and calibration tasks on virtual testbeds and XiL environments. The RD model can also be used in combination with the catalyst layer and membrane degradation models offered by AVL CRUISE M.

Example Use Case

Water management (i.e., the proper control of the membrane humidity) is a key aspect for PEM fuel cells. Humidifiers are often applied to use the humidity of the exhaust gas to pre-condition the inlet media streams. This is a typical use case for our electrochemical RD stack model in combination with the physical based humidifier. How does this work?

First, the PEM fuel cell system model, which consists of the electrochemical RD stack model, the anode/cathode media supply and thermal management systems, is calibrated based on high-fidelity 3D multi-physics component simulation result data and/or based on component/system measurements. Then, in a computational study, the humidifier configuration is varied together with other operating parameters (e.g., anode inlet humidity). For a single load point, we then get a trade-off between anode and cathode humidification. When the membrane dries out, this leads to a high ohmic losses, resulting in low proton transport and bad efficiency. Likewise, at high humidity, transport losses rise due to air dilution and eventually flooding, which again lowers efficiency. Optimizing this can be done independently for any steady-state and transient operating condition.

Now what does this mean for a vehicle under real-world driving conditions?

Considering the transient power demands and the driving resistances applied during real-driving conditions, the amount of hydrogen consumption and the driving range can both be predicted. This allows you to balance the fuel cell and battery sizing as well as optimize the powertrain system operating strategy.