Control Issues in Fuel-Cell Based Vehicles
L. Guzzella
Professor, Swiss Federal Institute of Technology (ETH)
Zurich, Switzerland
The intention of this talk is to introduce control people to the potential
of fuel-cell (FC) based vehicles and to indicate possible challenges for
modeling and control in this area. The talk will be structured as follows:
-
Fuel economy of FC based vehicles: The fuel economy of FC based
vehicles is very much dependent of the complete fuel cycle ("well-to-wheel").
Tools and data needed for this assessment and some worked-out examples
are presented in this part.
-
Some remarks on the electrochemistry of FC: Some elementary concepts
of FC electrochemistry are necessary to understand and to model especially
the static behavior of FC systems. Concepts like total free energy (Gibb's
potential), electrochemical limitations (activation and diffusion losses,
transport phenomena, etc.) will therefore be introduced in this part.
-
Modeling and control of FC systems ("stacks"): This section shows
what dynamic effects are relevant in FC systems and what control loops
must be closed for correct function. Also some important auxiliary devices
like hydrogen feed system and air supercharger are introduced.
-
Modeling and control of FC vehicles: FC based vehicles are very
similar to EVs, but there are some substantial differences: usually a "peak-shaving"
component has to be included (typically "supercapacitors"), when on-board
hydrogen reforming is chosen an additional "chemical reactor" has to be
included and the fuel storage part needs also some extra attention. The
hierarchical modeling, optimization and control problems of such vehicles
will be also addressed.
-
Ongoing projects at ETH/PSI: ETH/PSI has built a 1 kW FC system
that includes some critical components (stack, super-caps, etc.) and emulates
the remaining part through HIL simulations. In an ongoing project this
system is to be scaled-up to a 5 kW system with more components in the
loop (only the vehicle will be emulated). Some experimental data of both
systems will be shown.