Does anyone know what makes this more compelling than static power generation over the electric grid? Gas-fired power plants are pretty efficient. Is this significantly better?
Commercial natural gas electrical generation facilities are about 35% efficient. It is probably inappropriate to consider transmission losses since natural gas power plants tend to be located in areas of high consumption, but that 35% only goes down on the way to the end user. Remember, it doesn't have to beat the utilities for it to make sense.
Consider a house with a $200/month natural gas bill for heating. At $1/therm that is 200 therms of gas. Let's just guess at around 10% efficient for these cells, 220 therms of gas would give us our 200 therms of heat plus 20 therms of electricity which we convert to kilowatt hours, about 600 kilowatt hours. At $0.12 per kwHr that gets us about $72 of electricity for $20 in gas. Let's just say for every $3 we currently spend on gas we can get $1 of 100% efficiently generated electricity (1/3 the carbon footprint).
At 37°N, in December it would cover half of my electricity for a free standing house (but remember I made up their efficiency. At 20% it would cover all of my electricity. Googling about shows a record high around 60% and some press releases around 50%. I have no idea what these are, but it is probably several times the 10% number I used.)
During summer when I'm using nearly twice the electricity my gas bill is so close to zero as to not matter.
Sorry, this is wrong.
FCs are not heat engines, so the thermodynamic limitations of heat engines are not relevant to fuel cells. That said, thermodynamic limitations do set an upper bound on the efficiency of fuel cells. For hydrogen fuel cells, the theoretical upper bound is 83% [1]. That's for just the process of turning hydrogen into electricity, and that does not include the process for producing hydrogen in the first place. "90+%" efficiency is strictly impossible in theory or practice.
> i think the carnot efficiency maxes out somwhere 40-50% so no matter what you do (and we've been doing this for 100 years now)
It must be noted that 50% thermal efficiency is a practical upper bound for a real electrical power plant, not a theoretical one. Some combined-cycle gas turbines do exceed 50% efficiency under some operating conditions.
[1] http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/electrol.h...
1. Electricity grids have to account for a loss factor due to the accumulated resistance of the wires between the power station and where the electricity is actually used. This power is lost as heat radiating from the transmission wires. That said, I'm unfamiliar with gas transmission networks so I would assume there is some degree of potential energy loss through gas leaking from pipelines, etc. I expect it is lower than electricity grids however?
2. If the power generation is decentralised, I would assume that less would be spent on electricity substations, although that might be countered with a greater spend on maintaining the gas distribution network. I'm not an engineer, but I get the feeling that perhaps maintaining the pressure of gas to balance fluctuating demand is easier than maintaining the voltage/frequency of AC in a grid?
2.
EIA estimates that national electricity transmission and distribution losses average about 6% of the electricity that is transmitted and distributed in the United States each year.
http://www.eia.gov/tools/faqs/faq.cfm?id=105&t=3
Compare that to more than 50% thermal loss at the point of generation.