I'm going to make a WAG here that, when sourced from some kind of hydrolysis (there are various ways to do it) the electric-power-billed to at-the-motor-input for hydrogen-fuel cell is roughly twice as expensive (twice as much power used (ignoring the one-time capital cost (not in the operating budget) of the generation and storage equipment and assuming vehicle cost the same) as battery. At that level, it's certainly a niche thing, but the ability to refuel a rail vehicle in 1/2 hour or less without requiring the huge (and expensive) power supply to charge large batteries quickly is useful and operationally can offset the higher cost. Yes, the battery charging issue can be finagled using a large on-site battery and/or in-service charging at major stops and layover points, but there's a significant capital expense and the need for a large power supply (just not *as* large) remains. Batteries still take longer to charge than moving a similar amount of stored energy via hydrogen into tanks.

It's about twice the cost because the hydrogen fuel cycle (production-compression-storage-fuel cell) is about half as efficient as batteries. There are ways to manage that, chiefly use of renewable energy when available at the rate it's available (it's variable of course), storing the hydrogen short-medium term until it's needed. Really, it's not rocket science, but it's not a Big Business that can be dominated by a few colluding oil (or power, if on-site renewables can do the job) companies, and it's not a silver bullet (batteries and actual trolley wire can be more cost-effective in many cases), but it's useful in the right circumstances. Running a few FLIRTs for an operating day, refueling during daily servicing from hydrogen produced from solar and wind power while the trains are in use, is potentially cost-effective. It has worked in some bus applications as BOB2 has noted.

The ridiculous prices for hydrogen that have been in the press recently are really an artifact of industrial hydrogen mainly being produced from natural gas, and n.g. has been expensive lately, and truckers are hard to find and therefore expensive to haul it to filling stations. Throw in the almost negligible use of it for private cars and Big Oil rightly sees no future in retail hydrogen and appears to be trying to price it out of existence even if they don't just abandon the business like Shell did. Those issues don't apply to a properly designed and operated transit line mostly using renewable power to generate hydrogen.