I posted an article "A Hydrogen Economy - Is it Economic?" in January 2006, and concluded that we would need 180,000 2 MW wind turbines or 61 new Sizewell B capacity nuclear reactors to produce sufficient hydrogen to replace all the oil-based fuel used annually in the U.K. What I did not do was to factor in inefficiencies in the use of oil, and so I now revisit the concept bearing this in mind, and arrive at far more encouraging (though still daunting) numbers than these. It is a popular misconception that hydrogen is a fuel. It is not. It is an energy carrier. This means that hydrogen cannot simply be pulled out of the ground like gas, oil or coal, but it must be manufactured in some way. Most of the hydrogen produced in the world (used mainly for making chemical fertilisers for agriculture) is made by "steam reforming" natural gas, which is principally methane, according to the equation:
CH4 + H2O --> CO + 3H2. An "extra" portion of hydrogen can be squeezed-out of the system, by an adaptation of the water gas shift reaction: CO + H2O --> CO2 + H2, and so the overall process may be represented as:
CH4 + 2 H2O --> CO2 + 4H2.
The production of CO2 is naturally undesirable since it is a greenhouse gas, and so ideally a "green" source of hydrogen is wanted, e.g. electrolysing water using electricity generated using a renewable source like wind-power. Now, my question is, how feasible is this in terms of the generating capacity required to produce enough electricity to meet the scale necessary? Currently, we burn the equivalent of 57 million tonnes of oil each year to run the U.K. national fleet of vehicles, including planes (which consume around a quarter of that total, or 13 million tonnes). This leaves 44 million tonnes for road transport.
Fuel used in conventional internal combustion engines is burned very inefficiently, such that around just 16% of its total energy is recovered in tank-to wheel miles. Gas-Hybrid vehicles are far more efficient, and e.g. the Prius is reckoned at 37% tank to wheel. hence we could cut that total oil-bill down to (16%/37%) x 44 = 19 million tonnes of oil. Aviation is a separate issue, and the present calculation refers to road vehicles, because hydrogen-powered planes are very much a concept for the future, if ever).
Even if H2 could be provided on a large scale it can't be used with 100% efficiency either, and I shall assume an efficiency of 70% for the water electrolysis step, and 50% efficiency for an on-board fuel-cell, so that is the tank to wheel efficiency. This gives 70% x 50% = 35 % overall, in converting the electrons to road miles via hydrogen as the energy carrier. It is arguable that this is very inefficient to convert one form of energy carrier to another (electrons to hydrogen) and it is, but it is thought easier to store hydrogen than electrons, until better "battery technology" is developed. However that 35% is close enough to the 37% efficiency estimated for a gas-hybrid "Prius" vehicle that I shall compare oil with H2 on a one for one basis, using the 19 million tonne oil figure that would be required rather than the 44 million tonnes that we currently pour into our gas-guzzling internal combustion engines.
1 tonne of oil = 42 GJ of energy, and 1 kWh = 3.6 MJ. Therefore, 1 tonne of oil = 42 GJ/3.6 MJ = 11,667 kWh.
So, 19 million tonnes of oil = 19 x 10*6 x 11,667 = 2.22 x 10*11 kWh.
Now that's per year = 8760 hours, and hence the generating capacity = 2.22 x 10*11/8760 = 25,342 MW.
Let's look at two ways to generate this electricity: (1) nuclear and (2) wind power.
(1) Sizewell B has an electricity generating capacity of 1188 MW (the thermal capacity is nearer 3,600 MW, and so that inefficiency of converting heat to electrons has already been factored in). Hence we need 25,342 MW/1188 MW = 21 new reactors of this capacity to make the hydrogen to run our road fleet... and this is on top of the 30 or so new nuclear reactors that are needed by 2025 to replace the current nuclear generation which should be decommissioned... or mothballed by then.
(2) Wind turbines will need to be located offshore, in order to use the larger 2MW version with their 80 m long blade which is unpopular on land, for reasons of noise and spoiling the view. More practically, a greater capacity factor is obtained in offshore locations of around 0.4 as opposed to 0.2 for land based sites. i.e. each "2 MW" turbine would give an average of 0.4 x 2 = 0.8 MW. Hence we would need 25,342/0.8 = 31,678 of them, which is down considerably from my original estimate of 180,000, but is still a hell of a lot.
Now, the question remains of where would they go, precisely? I am making a very rough estimate, that the U.K. mainland can be approximated by an oblong 600 miles in length and 200 miles in breadth, giving a coastline of 600 + 200 + 600 + 200 = 1600 miles = 2560 km.
If we put them 0.5 km apart (which is the recommended separation) we could fit 2 x 2560 = 5120 in a single band. So, we need the actual band to be 31,678/5120 = 6(.19) turbines deep, and if they are 0.5 km apart, the band is around 3 km thick.
If the turbines were of 5 MW capacity not 2 MW (there are prototypes of this size) we'd need just 31,678/(5 MW/2 MW) = 12,671 of them /5120 = 2.5 deep, on average, and so the "band" would then need to be of 2-3 turbines on average and would present a thickness of about 1 km or so.
Along with all the problems of storing hydrogen, which needs high pressures and cryogenic cooling, even if it is adsorbed in porous materials like zeolites, and the fact that it make metal brittle and hence leaky (NOT good for an explosive gas), it might be better to store the electrons in "batteries" to get a better overall efficiency of 50% (70% of 70%) than 35% (70% of 50%) for hydrogen, when we could get away with around 20,000 2 MW or 8,000 5 MW turbines.
If we were to localise our society and cut transport by 90% we would be down to just 10% of this, needing only 2000 2 MW or 800 5 MW turbines. There would be no planes though, and if we want to keep them flying some other means must be found to do so.