For the pistons, ten to the tenth cycles and dynamic impact cycles at that. Who has reliable life data for that set of conditions and for what materials? To get to your system MBTF, wouldn't you need even higher cycle ratings?
Still, it would be a cool thing to work on, wouldn't it?

chuck,

That is about what I worked out for a year of continuous operation. What you need to do is keep the peak stress well below the fatigue limit. Which is probably why they need 200 pistons and not 2 or 20.

http://en.wikipedia.org/wiki/Fatigue_limit

200 cylinders * 1/second * 3600 sec/hr * 24hr/day * 365 day/yr = 6,307,200,000 to have a good chance of no failure during that time the cycle limit of a piston and its associated electronics/mechanical controls should be on the 60 billion cycles range. To prove that you would need 20,000 pistons operating for a month. Or 10,000 for two months. Or 5,000 for four months. At ~100 MW for 200 pistons that is a lot of energy.

Which may explain why - even though the machines are "cheap" - it will take about a billion dollars to develop a working machine.

M.Simon

Because the operation involves impact and (probably) high mechanical strain rates, if I were writing the check, I'd like to see testing to the target number of cycles at the projected peak strain rate and strain magnitude during the impact. Generally, the strain rate dependency helps, but for a guy like me, a billion still seems like real money. Getting the piston material sorted may be the least difficult thing they are attempting.

The fact that the guys involved used to design ink jet printer heads gives me a fair amount of confidence in their design.

In many ways what they are doing is making a very large ink jet printer.

My concern is their ability to hold timing.