[pfdietz]

That's grossly excessive. This website lets you look at optimization vs. actual historical weather data and reaches a much lower cost. BTW, you use both batteries AND storage; their combination can be cheaper than either alone. You also store the hydrogen underground rather than above ground.

https://model.energy/

[trashtester]

I agree that a combination of batteries and hydrogen would be a bit cheaper, but generally I would trust the MIT study over the model above, that states directly that it is a toy model.

It's fun to play with, though, so definitely upvoted.

Do you have an alternative peer reviewed study to support the conclusions?

[pfdietz]

(Edited to actually look at the link)

Looking at that first link, it appears to cost/kWh is the cost for just the seasonal power that is being produced from hydrogen. It is not the average cost/kWh over all the energy being consumed off the grid.

Comparing this to nuclear is to compare apples and oranges. Nuclear supplies baseload, not additional seasonal power. If you tried to use nuclear to supply seasonal power the cost would be extreme. If you use nuclear to supply baseload, it is competing not just with hydrogen but also with much lower cost energy sent directly from renewables to the grid, and from batteries to the grid.

Of course the fraction of the output from the hydrogen fueled turbines will be expensive. But what we get from that toy model is the average cost of power, of which seasonally stored power is just a small part. In that toy model, nuclear is "Dispatchable 2", which you can enable, and which comes out to about 10 euro/kWh (EPR assumptions, perhaps optimistic considering Flamanville 3.)

[trashtester]

About using nuclear to cover seasonal variation, its not ideal, but at worst you increase the cost by 50-100%.

Day/night cycles can make that even slightly worse, but even small amounts of storage can even that out.

This Korean study claims that a at 3% dicount rate, the price the LCOE of their APR1400 reactors go from 36Won/kWh (a bit under 3cents) at 90% average utilization to 52Won/kWh (a bit over 4 cents) at 60% average utilization.

In both cases, the costs are very reasonable.

[trashtester]

> Looking at that first link, it appears to cost/kWh is the cost for just the seasonal power that is being produced from hydrogen.

I interpreted those numbers as average LCOE for the electricy supply, but I could be wrong.

Let's assume that you're right, and this only applies to maybe 25% of total energy needs. The averge price would still be increased by $1000/MWh, on top of the renewable costs themselves, grid costs, etc.

[pfdietz]

If you go to https://model.energy/ and solve for California, 2011 weather data, 2030 cost assumptions, hydrogen is contributing just 5.9 euro/MWh to the average cost of electricity (less than 10% of the total cost of 68.1 euro/MWh). This is for supplying "synthetic baseload", the best case for nuclear (for varying demand, renewables could only do better, since one could always just swap out the nuclear source with this synthetic source, providing an upper bound on the cost.)

Nuclear will do best in more northerly places that are away from coasts. Poland, for example. For extremely northerly places, like Alaska, the total demand is so small that nuclear is not a good fit.

[trashtester]

> hydrogen is contributing just 5.9 euro/MWh

I'm afraid I don't trust those numbers, or rather, the calculations. For instance, if I select only Wind + Battery, and keep raising the storage cost for battery, the battery cost of the calculations go down, and are replaced by 100% wind plant costs. (Battery capacity is only enough for 1 hour of storage in such a scenario.)

On the other hand, if I drop the cost of battery to almost zero, the battery fraction of production goes to about 30-40%, and storage capacity is about 20 days.

In other words, it seems that the model is optimizing for an extreme overprovisioning of windmills when storage capacity is expensive, compared to when storage is cheap. (Basically, it "constructs" 10x or more windmills than what is needed from a pure energy production perspective.) I suspect that there are some issues with doing that is not taken into account by the model, such as available space at that level of density, risk calculations for completely dead winds, etc.

> Nuclear will do best in more northerly places that are away from coasts. Poland, for example.

Solar in dry, sunny places will likely be able to fully replace fossil fuels way before wind will be, since you do get some sunlight every day, while the wind may be gone for a week at a time. (And you have to plan for the worst case, in terms of available storage.)

[ncmncm]

No, because if wind is predicted to die down for a week, you order liquified ammonia shipped in from solar farms in the tropics. You might even keep a tank of a week or two's worth, banked against the occasion.

You may object that this seems fragile, but it is in fact how we live now. People are importing fuel from dodgy countries all the damn time.

[trashtester]

I alredy provided the link that led to the original paper, here is the original paper, though:

https://www.sciencedirect.com/science/article/abs/pii/S03062...