From Fuel Cells to a Hydrogen-based Economy

How vehicle design is crucial to a new energy infrastructure.
Fortnightly Magazine - February 15 2001

1 For example, a 1992 A.D. Little study estimated that a from-scratch bulk hydrogen supply infrastructure sufficient for 25 million cars would require about $95 billion of investment, or $3,800 per car. This antiquated result is still being quoted, e.g. in the Epyx article in the December 1998 (Derby 1998).

2 e.g., Lomax et al. 1997.

3 Widely quoted efficiency figures around 30-odd percent to 50 percent assume the fuel cell is fed not pure hydrogen, but the more dilute and impure reformate gas converted from a hydrocarbon fuel, and often include the conversion losses in the fuel processor.

4 Obviously, liquid fuels would become potentially interesting reformer feedstocks only if natural gas were not locally available, so that (for example) LPG or biofuels had to be substituted.

5 For illustration, even an $800 per kilowatt fuel cell system, at a 15 percent per year fixed charge rate, would incur a capital charge of only 2.7 cents per kilowatt-hour at a 50 percent capacity factor. Alternatively, the net electrical output efficiency of a PEM fuel cell using reformed methane often is quoted at or above 40 percent (lhv), often with neither heat recovery from the stack to the reformer nor pressure recovery from the stack's hydrogen input and stack output to the air compressor. With both forms of heat recovery, the best technology is now about 50 percent. At 50 percent conversion efficiency, natural gas at $3.70 per gigajoule or $4 per thousand cubic feet would produce electricity at 5.5 cents per kilowatt-hour. That would represent 2.7 cents per kilowatt-hour for the fuel plus 2.7 cents per kilowatt-hour for the cost of a relatively expensive early fuel cell system at about $800 per kilowatt, plus a nominal 0.1 cents per kilowatt-hour for O&M. This would undercut typical commercial-sector U.S. electricity tariffs (averaging 7.6 cents per kilowatt-hour in 1997) by 28 percent, even with no thermal credit and no allowance for the improved power quality and reliability or for other distributed benefits.

6 Lovins & Lehmann 1999, representing the capital and operating costs and the losses of the transmission and distribution systems for the average customer at the average hour. Obviously the actual costs, both total and marginal, depend on who, where, and when.

7 Lovins & Lehmann 1999.

8 Lenssen 1995.

9 Lovins & Lehmann 1999.

10 One-hundred fifty million light vehicles times a minimum capacity of 20 kW-the average could be substantially higher-yields 3 terrawatts (TW), vs. summer-1997 U.S. peak capability of 0.78 TW and 1996 noncoincident peak load of 0.62 TW (neither of which reflects the approximately 14 percent on-peak grid loss).

11 Bain, 1997; Bain, Addison: Personal communication, Nov. 1, 1999.

12 Directed Technologies Inc. 1997.

13 James et al. 1997. Further, a fuel cell Hypercar could travel roughly 200 km on 1 kg of hydrogen: A Taurus-class Hypercar was calculated to drive roughly 925 km fueled by 4.65 kg of hydrogen (Williams et al. 1997).

14 President's Council of Advisors on Science and Technology (PCAST) 1997 at 6-34.

15 Ogden et al. 1997, Thomas et al. 1997, 1998a. Although natural gas reformation generally is assumed to be the cheapest option, if off-peak retail electricity costs only about 1.5-3 cents per kilowatt-hour, as it now does in many parts of the nation, using it to split water could cost less than locally reforming natural gas for small numbers of fuel cell vehicles: Thomas et al. 1998, 1998a. Electrolysis could therefore initially be deployed faster if initial vehicular hydrogen markets were small, but vehicle fleets or, of course, fuel cell systems in buildings could favor small methane-steam reformers.

16 Thomas et al. 1998a.


18 Lovins 1998. On-site storage of compressed hydrogen is straightforward, although updating of regulations is necessary. In general, current regulations, meant for natural gas, assume metal tanks subject to corrosion and cracking, and are overly conservative for the very different engineering details of composite hydrogen tanks.

19 Such hydrogen appliances could even end up in individual garages-better than schlepping one around in your car, and providing battery cars with an overnight refueling advantage. Electrolyzer Corp. of Canada, for example, is developing just such an electrolyzer-and-compressor device for home use.

20 Assuming 18,000 stations each able to supply 1,000 relatively conventional (about 40-kW to 80-kW) fuel-cell vehicles-an offboard investment of $230 per vehicle: Thomas et al. 1998.

21 Williams, R.H. 1996.

22 One gallon priced at $1.25 contains 125 kBTU or 132 megajoule, enthalpically equivalent to 36.6 kWh of electricity priced at 3.4 cents per kilowatt-hour. However, the 2.5-3.5-fold greater efficiency of converting each J of hydrogen into vehicular traction, compared with a J of gasoline (i.e., about 50 percent system efficiency in a fuel cell car vs. 15 percent to 20 percent in an Otto-engine car), makes this price functionally equivalent to 8.5 cents to 12 cents per kilowatt-hour.

23 This is bad news for aluminum smelters, which now often enjoy preferential access to very cheap hydropower under old contracts that will represent an increasingly severe opportunity cost. However, it might be good news for anadromous fish if hydrogen storage could be large enough to control seasonal water flows for their benefit.

24 Only the road-vehicle portion of transportation's emissions, of course, but fuel cells in buildings and industry would also displace much of the fossil fuel now burned for space, water, and process heating.

25 Lovins 1998.

26 If one multiplies the per-vehicle costs in Thomas et al. 1998a times the world's fleet of a half-billion light vehicles, which is growing by about 5 percent per year.