William Burton asked an interesting question in the comments on my first fuel cell posting. While I addressed it there, it did make me curious. Could the ancillary benefits of fuel cells outweight the costs? I decided to do a little research, followed by some low rent modeling to see exactly where things fell out.
Let's start with the IC engine. As technology has improved, we have come a long way in IC engine efficiency. Today, a standard engine runs in the neighborhood of 32% efficient, not counting drive train and friction losses, which will be the same regardless of power plant design.
Obviously, this is far worse than the 85% efficiency achieved by the fuel cell. The problem comes when we fuel the cell. Making the hydrogen requires cracking water, and the standard method involves electricity. A steam generating plant, whether fired by coal, fuel oil, biomass or some other fuel typically runs at about 34-38% efficient. So combining the steam plant with the fuel cell efficiencies gives us a range of 28-32%. So, in the best case, we merely equal the efficiency we already achieve, at a greatly increased cost. We're using the same amount of fuel, or better, resulting in the same environmental impact, the same dependence on foreign oil, only we're spending a lot more money to do it.
Now, let's examine solar power for a moment. I did a quick search on the net, and the best efficiencies I could find for solar voltaic conversion was 28.7%, and that was using a terrestrial version of solar modules designed for use on satellites. The cell yielded 2.571V at 12.95mA/cm2 on a 30cm2 area. A standard solar cell usually runs at around 12% efficient, giving 10mW at .6V on a 2 cm2 cell.
Time to pull out our trusty conversion calculator. We'll assume we have economical access to the super cell.
Power=Voltage X Current=2.6V X 12.95mA/cm2 = 33.7 mW per cm2, or 1010mW per cell.
Let's assume a typical power plant size of 700MW.
700MW/1010mW per cell=693 million cells or 746,000 square feet, or just over 17 acres of solar panels. Except it will probably be more like 21 acres, if you figure 20% overhead for operating equipment, framing materials, access paths for maintenance, etc.
My trusty Pocket Ref tells me that 1 horsepower is equal to 746 W so this 700MW power plant will produce 940,000 horsepower. Now to make things easier, we will neglect conversion losses and assume that all available power is transformed from electricity to hydrogen. The true efficiency is probably around 95%, so we aren't introducing much error. However, we do have to account for the 85% efficiency of our fuel cell, which reduces our available horsepower to 800,000. Next, we'll assume that the auto fuel cell is around 100 horsepower. We'll also assume utilization at 10% since cars do sit idle most of the time, giving us an effective horsepower of 10. Now we have a grand total of 80,000 fuel cells per 700MW power plant.
Now, at any given time, there are approximately 175 million cars in the hands of our get up and go population, meaning we will need almost 2200 new solar power plants just for the auto industry alone. 2200 plants at 22 acres per plant, add in another 15 acres (very conservative) for physical plant, parking etc, and you have 48400acres, or 76 square miles of solar panels.
Next we have to consider the costs of constructing a delivery system for the hydrogen, one which is efficient, and safe. Liquids can be piped relatively easily. Designing a nation wide network of pipes to contain hydrogen proves to be trickier, and significantly more expensive.
So, at a rough glance, in order to replace the IC engine with fuel cells, we will need to construct 2200 new power plants, whether solar or conventional, and develop a nationwide delivery system for a very volatile gas. As SKB suggested, there are stabler forms of hydrogen, but none of those forms are suitable for fuel cell use, except methanol, which involves CO2 emissions, negating the chief advantage to the fuel cell. As far as I can tell, the benefits are minimal, compared to the expense.