Newfie wrote:Real, Do you know what the
energy conversion effiencies are? Theoretical vs current?
Newf, the conversion efficiency was once poor but there are many new methods. Here are the notes I have on hydrogen from various forums I visit. I apologize for the length but this is a very important topic worth a look:
1. Concerning the cost of electrolysis infrastructure. The most up to date cost estimate for electrolysis comes from the ‘International Journal of Hydrogen Energy’ and is titled: ‘Future cost and performance of water electrolysis: An expert elicitation study’.
The paper produces cost projection for three electrolysis technologies at full scale production.
Technology 1: Alkali Electrolysis Cells. Full scale unit cost averaged at $1400/kWe in 2015 and was projected to be $1100/kWe by 2030.
Technology 2: Proton Exchange Membrane Electrolysis Cells. Full scale unit cost averaged at $2100/kWe in 2015 and was projected to be $800/kWe by 2030.
Technology 3: Solid Oxide Electrolysis Cells. Full scale unit cost averaged at $5800/kWe in 2015 and was projected to be $1100/kWe by 2030.
Whilst it is difficult to predict exactly where these costs will be in 10 years’ time, these estimates suggest that Alkali electrolysis can already deliver capital cost of $1400/KWe and that a cost of $1000/kWe is a reasonable assumption for one of the three technologies considered. Is Cloggie’s estimate of $500/kWe possible? Yes –with greater than expected cost reductions and large economy of scale driving down stack costs. But I think this should be considered a stretch case, rather than something that is automatically expected. For the time being, I will work with $1000/kWe, because it is the value supported by peer reviewed study.
2. The estimated 2030 lifetime of electrolysis cells is 60,000-90,000 hours (AEC); 20,000-90,000 hours (PEMEC); 30,000-90,000 hours (SOEC). I am going to assume a median value of 75,000 hours for SOEC. This translates to non-stop lifetime of 8.6 years, or 28.5 years at 30% capacity factor, assuming that thermal cycling does not result in unexpected damage. I am going to assume an interest rate of 5% on borrowed money and a flat repayment rate of $68.5/kWe, which allows complete payoff of capital cost in 27 years.
3. Stack efficiency. This includes losses to things like pumps, heaters and water purification, so it is a whole system, rather than simple cell efficiency. For AEC: 50-75%; for PEMEC: 50-78%. Pumping losses and thermal losses will all be minimised in systems scaled to hundreds of MW, so I will assume efficiency of 70% in a large scale, well optimised system.
4. Stack capacity factor. This depends heavily upon the assumptions made for the driving power source. An electrolysis plant using solar power in the desert will be in darkness ~50% of the time and illumination will follow a sinusoidal pattern between sunrise and sunset. I am going to assume a capacity factor of 30% and further assume that the liquefaction plant can absorb peaks in power supply by storing thermal energy in an intermediate cold store.
Putting these figures and assumptions together, an estimate can be made for the capital repayment costs associated with the electrolysis unit. At 70% efficiency, to produce 1kW hydrogen would require 1.43kWe cell capacity. The repayment costs are $97.86/year. At 30% capacity factor, the 1.43kWe cell will generate some 2630kWh of hydrogen per year. So the marginal capital cost per unit of hydrogen energy is 3.72 Euro-cent per kWh.
If the capacity factor can be increased to 50%, say, by using a mixture of renewable energy sources to provide the driver current, then marginal capital cost drops to 2.23 Euro-cent per kWh hydrogen. At 90% capacity factor, driven by a nuclear reactor, marginal capital cost declines to 1.24 Euro-cent per kWh hydrogen.
Absent from this calculation is any consideration of operations costs, electricity cost or profit. And of course we have not considered the energy or operations cost of any liquefaction plant. Ignoring liquefaction and assuming operating costs to be about 1/3rd of capital costs, some estimates can be made.
Solar powered scenario: Lazard (2018) provides a low-end LCOE for thin-film solar at $40/MWh. Given that we will probably choose to locate a hydrogen plant where there is plentiful sunshine, this value will be assumed. Some 1.42kWh of electricity are needed to produce 1kWh of hydrogen. This puts the electricity cost of 1kWh of hydrogen at 5.68 Euro-cent. The final LCOE for 1kWh of hydrogen gas produced using thin film solar is therefore: 1.33 x 3.72 + 5.68 = 10.64 Euro-cent per kWh.
The costs associated with a liquefaction (or other chemical conversion) and / or storage facility, will be additive to this.
My guess is that imported liquid hydrogen will be used for large scale power production in combined cycle gas turbines, much as LNG is today. LNG has not caught on as a bulk vehicle fuel, although it is both cheaper and much easier to handle and distribute than liquid hydrogen. Why would we expect hydrogen to be used in this way when LNG is not? There are solid oxide fuel cells that burn methane quite efficiently.
For bulk applications like power production, the question for a European operator will be: why use more expensive liquid hydrogen, when I could use cheaper LNG? For a power plant operator, fuel cost subtracts directly from their bottom line.
“Hydrogen will only be used as a brief intermediary stage and almost immediately converted in another more convenient storage medium: NaBH4, CH4, NH3, metal powder”
Methanol is another promising option that you have mentioned before and is probably better than any of the others on the list, due to its excellent properties. It is reasonably energy dense (about half as energy dense as diesel), is storable as liquid at room temperature at atmospheric pressure; will not freeze until -97C and is compatible with carbon steel. All winning combinations that make it very useful for end use applications.
There are fuel cells that can use methanol directly, that can be scaled to power tiny individual devices, like laptops. It can be used in spark ignition and compression ignition engines and would work well in the hybrid vehicles of the future. It could also be stored easily long-term in steel or concrete tanks for back-up powerplants; things like open-cycle gas turbines, which need to cover occasional periods of peak demand, or where renewable energy and short-term pumped storage occasionally fails to meet grid demand. In the home, a direct methanol fuel cell or small gas turbine could provide combined heat and power.
If you have cheap hydrogen from renewable energy, then any carbon containing material could be decomposed to produce the carbon monoxide needed to synthesise methanol. This might include low grade fossil fuels, like lignite or kerogen; wastes that would otherwise go to land fill; biomass; sewage, animal wastes, marine sediment, etc. Or it might include the carbon dioxide waste from a concrete kiln.
Using limestone as feedstock is an interesting option. The calcium hydroxide that comes out the other end could have uses of its own, as lime for building or for neutralising acid soil. Or it could be released into the ocean in areas suffering from ocean acidification (it is a base afterall). As it neutralises carbonic acid in the ocean, it will precipitate to the bottom as a carbonate. Limestone derived methanol would therefore be carbon neutral, or even carbon negative, if some of the liberated carbon is fixed into products like plastics.
Boron would appear to be a workable option for long-term hydrogen storage. It has a few disadvantages. Boron is relatively rare and expensive – it must be collected after use and transported back to the hydrogen plant for recycling; sodium borohydride requires a chain of chemical processes in its manufacture (each of which consume energy); and so far as I know it is solid and cannot be pumped. Iron has similar issues. It also needs to be burned in a large boiler. Anhydrous ammonia is a carbon free fuel; nitrogen is available everywhere from the atmosphere and is storable as liquid under modest vapour pressure. But it is smelly, irritant and toxic, which will complicate its use as a fuel.
I have a copy of Olah’s book – Beyond Oil and Gas: The Methanol Economy. It is one of the better alternative energy books out there in my opinion. In the past, the high cost of electrolytic hydrogen was an obstacle that prevented the realisation of a methanol economy. Not so much today.
Another thing paving the way towards a methanol economy is an often forgotten by-product of electrolysis: pure oxygen. During electrolysis, a steady stream of compressed pure oxygen forms at the anode of the cell. In an oxidation reactor, this will break down any organic material at very high temperatures, far more effectively than ordinary air. Even wet biomass can be converted into synthesis gas by reacting it with pure oxygen and steam – something that would be difficult to achieve using air. The hydrogen from the cathode will add to the hydrogen already produced by oxidation and steam reforming and can simply be added to the synthesis gas passing into the methanol reactor. Given that carbon dioxide can be reduced back to CO by reacting it with hydrogen, close to 100% of the carbon within biomass can be captured in liquid methanol.
“Add to that electrolysis at 0.6 cent/kWh, that is 3 cent. Add 1 cent for 80% electrolysis efficiency and you have 4 cent or €1,32 for a kilo of hydrogen.”
This is not realistic, even with dirt cheap solar power, which may not be sustainable. My previous analysis was based upon projected efficiency and cost improvements for 2030. It calculated that with solar electricity costing €0.03/kWh; hydrogen gas at the cathode of the electrolysis cell would cost €0.11/kWh. That is about 3-4 times the cost of diesel before tax.
By the way, 80% electrolysis stack efficiency is not realistic. This is the cell efficiency, before other losses associated with heaters and pumps within the stack are taken into account. When these are accounted for, efficiency is 60-70%, maybe 75% by 2030; time will tell. Here is my previous analysis; I have posted it here before.