Fuel cells: stack control too simple for the lifetime asked

Stack control has to fit on small hardware and stays simple

What happens. The fuel-cell system runs on a compact control unit with little computing power. Air supply, hydrogen pressure, humidity and coolant temperature are each held by a standard controller with lookup tables. The tables are filled on the test bench, operating point by operating point, and the interactions between the loops are covered by margins.

What it costs. Every new stack, compressor or cooling layout means weeks of filling tables again. Load steps cause short spells of too little air or too dry membranes, which the stack does not forgive. The margins cost efficiency over the whole operating range.

Why the usual fix fails. More tables and more correction terms make the software harder to validate and still only cover the conditions that were measured. A bigger control unit raises cost and power draw, and a model-based controller is often assumed to be too heavy for the hardware.

How a fixed-price remote diagnosis checks it. From bench measurements, the current control structure, the tables and the specification of the control unit I check how much of the hand tuning a compact model of the stack and its air path could replace, and whether such a controller fits the available computing time. The fixed-price remote diagnosis shows what it would take and where the limits are.

Background: Control on embedded hardware

A new fuel-cell or hydrogen powertrain needs its control concept from scratch

What happens. A team that knows combustion engines or battery drives starts on a fuel-cell or hydrogen powertrain. Stack, compressor, humidifier, hydrogen recirculation, battery and cooling have to work together, and the operating strategy for cold start, load changes and shutdown does not yet exist.

What it costs. Without a clear control concept, the first prototypes run with conservative settings and many manual interventions. Test-bench time is spent finding out how the subsystems interact instead of improving them. Late changes to the architecture delay the whole programme.

Why the usual fix fails. Carrying over the control structure of the old powertrain misses what is new: humidity, purge cycles, ageing of the membrane. Hiring a complete control team for a technology that is still being proved is slow and risky.

How a fixed-price remote diagnosis checks it. From the system layout, component data sheets, first bench measurements and the planned operating cycles I sketch which variables must be controlled, which can be measured and which must be computed, and where the subsystems interfere. The fixed-price remote diagnosis gives you a reviewed control concept to build on, with the open questions named.

Stacks must last for years, but heat and dry operation age them fast

What happens. Fuel cells for heavy-duty use must reach a lifetime of more than 25,000 hours, while high temperatures and low humidity increase their degradation (German Aerospace Center). In a truck or bus the stack sees hot days, long climbs and sudden load changes. The energy management decides how much the stack delivers and how much the battery covers.

What it costs. A stack that ages early is one of the most expensive parts to replace. Running it gently on fixed limits wastes efficiency and battery capacity; running it hard to save hydrogen shortens its life.

Why the usual fix fails. Fixed power limits and temperature margins protect the stack in the worst case but ignore how it is actually used. Ageing cannot be measured directly in operation, so the strategy is set once and rarely revisited.

How a fixed-price remote diagnosis checks it. From vehicle logs, stack voltages and temperatures, humidity data and the energy-management parameters I estimate how operating conditions relate to the observed voltage loss. The fixed-price remote diagnosis shows whether an energy management that weighs hydrogen use against ageing could work, and what data it needs.