Monday, March 11, 2013

Wind for Hydrogen - An Update.

This is an update of some numbers from an older posting http://ergobalance.blogspot.co.uk/2007/10/ulf-bossel-platinum-and-hydrogen.html in light of the larger commercial wind turbines that are now available http://ergobalance.blogspot.co.uk/2012/01/shaky-foundations-for-offshore-wind.html. There is some improvement in the overall figures, but the task of switching from oil to hydrogen remains stupendous.

At the outset, we should note that hydrogen does not occur free in nature but must be freed from other elements, such as oxygen in water, with which it is naturally combined, and the separation of elements requires other forms of energy. Almost all the hydrogen used currently in the world - principally as a chemical feedstock e.g. for oil refining and making artificial fertilizers - is made by steam-reforming natural gas, and there is a CO2 budget that must be costed-in, hence hydrogen from this source is not clean but contributes to CO2 emissions. Furthermore, it consumes natural gas, and so there is a further demand placed on another resource, in accord with the indisputable fact that it takes resources to extract resources. Ideally therefore, that hydrogen should be produced by e.g. water electrolysis using electricity made from renewable sources.

Some while ago, Ulf Bossel pointed out http://www.fuelcellforum.com/reports/E21.pdf there are losses at each stage in the chain of production, storage and distribution for hydrogen. There is obviously a loss of 50 - 60% incurred when the material is oxidised in a fuel cell, but in its favour is the fact that an efficiency of even 40 - 50% is substantially above the Carnot-cycle limit (Thermodynamics again) of around 35% for a typical internal combustion engine. The losses may be summarised as follows: 90% efficiency for rectifying alternating current to DC to run the electrolyzer; 75% overall efficiency (ideal) for the electrolyzer itself; and then the storage of the bulky hydrogen gas either as a highly compressed gas, which takes about 20% of the energy content of the hydrogen to compress it (or as a cryogenic liquid, which takes 30 - 40% to produce); 10% for distribution and say 50% efficiency for the fuel cell itself, which amounts to about a 25% efficiency overall.

There are electrolyzer units http://www.nrel.gov/docs/fy04osti/36705.pdfthat can produce high pressure hydrogen and if each gas-station were to make its own hydrogen by electrolysis, much of the distribution losses (probably 30%) might be avoided. A report has been published by a firm of independent analysts in Germany which is critical of some of Bossel's figures http://mpfc.de/pdf/LBSTonBossel.pdfespecially in regard to storage and transmission, particularly across large distances say from sunny north Africa (if the hydrogen were produced using PV technology which would be much more efficient there) by pipeline to Europe. However, an in-situ arrangement as I allude to would surely get around that, presuming we could make enough renewable electricity, or if there were a grid of electrons (rather than of hydrogen) including north African PV, European wind-power, North Sea wave energy and so on, such power might be supplied to run local electrolysis equipment, which would avoid actual hydrogen transmission. But if Bossel is right, why not use these electrons in a more direct manner?

On a tit-for tat basis, we can make the following calculation:

The heat of combustion of hydrogen is -285 kJ/mol, and so 1 kg of hydrogen = 1000 g/2 g/mol x -285 kJ= -142,500 kJ = 1.425 x 10^8 J.

We get through 82 million tonnes of oil altogether annually in the UK and we use 60 million tonnes of that for fuel. The energy content of oil is rated at 42 GJ/tonne and so that 60 million tonnes "contains" 60 x 10^6 x 42 x 10^9 Joules = 2.52 x 10^18 J of energy.

Hydrogen can be produced at a pressure of up to 10,000 psi by electrolysis at a rate of 60.5 kW/kg of H2. Hence the equivalent H2 to match that amount of oil is:

2.52 x 10^18 J/1.425 x 10^8 J/kg = 1.768 x 10^10 kg H2. Bossel has used the conversion factor of 1.5, i.e. that H2 can be used with 1.5 times the recoverable energy efficiency of gasoline. Since gasoline gives an approximately 14% well-to-wheel efficiency that would make about 21% overall for hydrogen, which seems a bit low and I would think that say 59% for the electrolysis system x 90% for rectification x 50% for the fuel cell = 26.6% is more like it.

However, let's consider the generating capacity the whole enterprise would need. To make 1.768 x 10^10 kg of H2 over a year, i.e. 8760 hours, would require:

1.768 x 10^10 kg x 60.5 x 10^3 (W/kg H2)/8760 = 122.1 GW. But this figure is mitigated according to the efficiency with which hydrogen may be used. If Bossel is right, this becomes 81.4 GW or let's call it a factor of two (which seems more reasonable), making it 61.0 GW.

Either way, we would need a colossal installation of renewables, e.g. 5 MW wind-turbines, with a rated capacity of 5 MW - but an actual output of say 30% if placed offshore, which amounts to 1.5 MW per unit. Hence we would need 61 GW/1.5 MW = 40,667 of them. Probably these could be accommodated in the North Sea in a 202 x 202 square of turbines, and at an average spacing of "ten rotor diameters", i.e.  1.23 km,we are talking about an area of 247 kilometers squared (= 61,113 km^2), which doesn't sound too bad, albeit that the weather in the North Sea is some of the roughest in the world, and so maintenance might prove a problem. If the turbines were placed around the coast of the U.K. mainland (assumed to be 2,500 km in length), at a mutual separation of 1.23 km, 2,033 turbines could be so accommodated in a single strand, and to contain all of them, a band would be created, 40,667/2,033 = 20 turbines deep. Assuming that same 1.23 km separation, this would be 25 km (15 miles) wide. So, how quickly might this farm of 40,000+ wind turbines be created? The question really is one of "how long is a piece of string?" but assuming that one turbine could be fabricated and installed every day, the process would take at least 111 years, and probably far longer, in reality.

As an alternative, around 60 new nuclear reactors could be installed to make the electricity for hydrogen, and on top of the new generation required to replace the decommissioned current 31 reactors, actually equal in output to about 14 1 GW reactors, and so it would be necessary to quadruple this capacity by which means to install a "Hydrogen Economy" in the UK. I have been told that hydrogen could be made more efficiently using the thermal power from a nuclear reactor to run the iodine-sulphur cycle, rather than by electrolyzing water (50% compared to 35%) , but the installation capacity needed remains huge. If Bossel is right and electrons can be used with three times the efficiency than will be recovered (hydrogen actually re-generates electrons in the fuel cell, to turn wheels, in a chemically-fuelled electric car) by turning them into hydrogen, the installation capacity immediately falls to 20 new nuclear power stations, or about 13,555 turbines, which is still enormous but appears more achievable.

I am not ruling out hydrogen altogether but simply making the point that when oil supplies begin to wane, it is not a simple matter of switching from oil to hydrogen, but a new and vast infrastructure must be implemented first, to both produce and use hydrogen. The question looms: is it worth it, or might there not be better ways to deal with our impending transportation problems, such as relocalising society to use less transport? Even those who are profound advocates of the "Hydrogen Economy" need to address the problem that the PEM (Proton Exchange Membrane) cell relies on an electrode consisting partly of platinum (about 50 - 100 g worth), which is a metal so rare than only 200 tonnes of new platinum are produced each year, and well below the current and growing demand for it.

Admittedly, the 40% of world platinum that is presently put into catalytic converters could be fabricated into PEM cells, were the putative conversion from oil-power to H2-power to be made, but this is only sufficient to put around: 200 tonnes x 1000 kg/tonne x 1000 g/kg x 0.4/50 g/cell = 1.6 million new "vehicles" on the road each year, out of a world total of about 1,000 million. Hence over a period of 15 years we could replace just 2.4% of the current number. Thus, unless more platinum is recovered on a huge scale (from sources as yet unknown to geology), or some alternative fuel cell technology is brought to a commercial level of development on some similarly immediate timescale, the enterprise looks set to fall at the last fence, in this, the last race that humankind will ever have to place bets on.

2 comments:

Trim said...

Chris,

More articles like this one please! Outstanding presentation of reality.

Cheers,

Ken

S/V Trim

Professor Chris Rhodes said...

Hi Ken,

I wrote quite a number of "quantitative" articles like this particularly back in the early days of this blog. Maybe it's time to revisit some of them (as I have done here), or/and once more apply the approach to some of the newer factors such as fracking, shale gas/oil etc.

Cheers,

Chris