Five areas, worked at system level.

Fuel-cell power systems and the balance of plant around them: the supply, the thermal path, and the interface to the rest of the vehicle.

A fuel-cell stack is a component. A fuel-cell power system is everything that has to exist around it before that component becomes usable on a real platform: reactant supply at the right pressure and temperature, a thermal path that removes heat where there is no room for a radiator, and the electrical and mechanical interfaces to everything else on board. We work at that level, inside the mass and volume budget the platform allows.

• Balance of plant
• Thermal path
• Interface to the rest of the vehicle

Machines and drives, and the interfaces that put their power into a shaft or a propeller, inside the torque, speed and cooling envelope the platform allows.

Electric machines and their drives are available as products. The engineering is in the fit: matching the torque and speed the platform actually asks for, mounting and cooling the machine where it has to sit, and getting power to it and motion out of it without introducing a failure the rest of the system cannot handle.

• Machines and drives
• Mechanical interfaces
• Electrical interfaces

Conversion, distribution and protection between sources, storage and loads, within the voltage and thermal limits of the platform. Faults are contained, not passed on.

This is where a system with two dissimilar sources becomes one system. Conversion and distribution between the power source, the storage and the loads, protection that contains a fault rather than propagating it, and thermal design that holds at the duty the platform will really see. Efficiency is won or lost here, and so is safety.

• Conversion
• Distribution
• Protection

Control and monitoring hardware and software, specified against the behaviour the system has to hold when something fails, not only when everything works.

Control and monitoring hardware and the software that runs on it: how the system is supervised, what it does at its limits, and how it behaves when a sensor, a source or a subsystem is lost. The specification that matters is the one written for the failure case, because that is the one that has to be demonstrated later.

• Control hardware
• Monitoring software
• Behaviour under fault

Rigs, instrumentation and test campaigns that establish on the ground what a system does before it is trusted in the air.

Rigs and instrumentation built to answer specific questions, test campaigns run against the requirements rather than against a hope, and verification evidence that is reviewed and documented. Prototypes are built with production intent from the outset, so that what is learned on the bench still applies to the thing that flies.

• Rigs and instrumentation
• Test campaigns
• Evidence before flight