In previous post, I wondered where exactly in the calculator methane storage was implemented and how seasonal storage was dealt with. Spoiler alert, no methane storage whatsoever is implemented in the calculator and seasonal storage is not taken into account. But then, how is the proposed Power-to-Methane-to-Power backup system implemented in the spreadsheet?
The way it works is that only the totals are used in the final cost formulas and his means that problems like intermittency and seasonal storage are ironed out of existence. I can surely understand that ON AVERAGE solar and wind can keep up with demand, but this is not how a power grid works. In reality, demand in ALL time slots need to be fulfilled by solar and wind. That is where the spreadsheet is lacking.
I was puzzled for a while why NIsche didn’t add storage in the calculator. Reading his earlier post ““Correcting Anti-Renewable Energy Propaganda” post” gave me some insight in his reasoning. In that post he explains his vision of a 100% renewable electricity grid and his reasoning is that overbuilding solar and wind capacity 1.5 times the demand would already get 93% of demand covered without any storage at all. The remaining 7% then just needs to filled in with synthetic methane.
Where that methane needs to be stored until it is burnt in gas-fired power plants is not really clear from this post. He only gave the example of a “facility that generates renewable methane and injects it into the gas grid”. I can understand that putting the resulting methane on the gas grid will give some buffer, but is that enough to store it in the necessary amounts?
Particularly revealing is Nitsche’s comment in the same post as a response to someone criticizing the Power-to-Methane-to-Power part of the reasoning (my emphasis)
Technically, if you define “grid storage” as “surplus electricity is stored and reused when there’s a lack of electricity”, that’s right, power-to-gas is not storage in that sense. You generate a high share of solar and wind, say 93 percent, in real time. You curtail surplus electricity instead of storing it.
In addition to that, using different solar and wind power plants, you generate synthetic methane (hydrogen + CO2-direct-air-capture). This methane is burned in gas-fired power plants to cover the remaining 7 percent.
This is indeed exactly what he did in the spreadsheet. The total production is 1.56 times the demand. However, the demand that is directly covered by solar and wind is even higher in the calculator. It is 97.5% in the spreadsheet, so only 2.5% of the demand needs to be filled in. I can be wrong, but it is my impression that Nitsche assumes that no dedicated storage is necessary for this small amount.
This might make perfect sense when you look at it on average, but is that the case in reality? It assumes that production and demand are pretty much in balance over the year and not much storage is needed to bridge the periods of deficit. Even in the averaged dataset that he uses, there is a clear pattern of how this deficit is distributed over the year (data of the last year of the dataset, 2018):
This shows the discrepancy between production and demand over a year. Production is high in the first five months of the year, but then decreases and reaches it lowest point in August, after which production climbs up again. Demand makes the opposite movement. It climbs up from May and reaches its highest point in August after which it decreases gradually until October. Meaning that there is a huge overproduction in the first five months and the last three months of the year, but the highest need for backup power occurs during the summer months when production is at its lowest.
| Month | Deficit (GWh) |
Share (%) |
|---|---|---|
| January | 367,881 | 9.3 |
| February | 243,580 | 6.1 |
| March | 104,759 | 2.6 |
| April | 27,702 | 0.7 |
| May | 99,086 | 2.5 |
| June | 252,155 | 6.4 |
| July | 891,133 | 22.5 |
| August | 1,029,653 | 26.0 |
| September | 371,198 | 9.4 |
| October | 123,086 | 3.1 |
| November | 125,020 | 3.2 |
| December | 328,824 | 8.3 |
| Total | 3,964,079 |
There is clearly a need for seasonal storage if one wants to use this overproduction from the first months of the year in order to fill in the deficits during the summer months. In an actual 100% renewable electricity grid, those seasonal differences are a problem that have to be solved using for example seasonal storage. In the calculator however, this seasonal variability is ironed out of existence. This means that all costs associated with (seasonal) storage are entirely ignored.
The averaging also means that no limit is assumed for the production of the methane used as backup fuel. The cost of the production is determined by the production cost of methane and multiplied by the surplus electricity, so no capacity is attached to it. However, that methane production cost is taken from a plant at full load, but in reality such factories will have a certain capacity and this will limit how much methane can be produced and how much could go into storage. Dimension it large enough to produce enough for the period June to September and it will not be that efficient. Dimension it to be more efficient will mean not enough methane in summer.
The same is true for the methane storage facilities. They are completely absent from the model, but in reality they will be needed to bridge seasonal shortages and they will also have a certain capacity attached to them that determines how efficient they will be. Seasonal storage is inefficient, therefore expensive.
Then there is the curtailment. The system is overbuild and electricity that could not be used, will be curtailed. That would make sense on average, but I don’t think that is how it would work in reality. Here, across the pond, industrial solar and wind installations are built by power companies or investors. The way they can recuperate their money is by selling electricity to the grid. Since they get a guaranteed price and some protection against negative prices, they can make a buck.
Things are different in the spreadsheet. Here a lot of solar and wind capacity is build, it is in fact overbuild to produce more than needed and the rest is curtailed. Meaning that the investors will get less income from selling their electricity to the grid and therefore their investment will pay off over a longer period. I am not really sure how such a system would work in reality.
Here in Belgium, we had a nice illustration of what would happen when solar and wind farms need to curtail their production. In the first month of the pandemic, electricity demand was pretty low and there was a lot of sun and wind. This meant that some windmills were shut down for some hours to avoid overproduction, leading to the CEO of an energy company to write an open letter to the politicians to “choose for solar and wind” instead of nuclear (read: shut down nuclear, so solar and wind could earn their buck when it is sunny and windy). The reason behind this was that those windmills were in fact shut down to avoid providing electricity at negative prices. Although those providers are protected against such negative prices, this support is not unlimited.
Now suppose that the grid was organized according to NItche’s plans, then we are looking at not just a couple hours of curtailment, but systemic curtailment of electricity, specially in early spring, autumn and winter). Remember, Nitche starts from a production of 1.5 times demand and only 2.5% of demand comes from methane. Which investor in his right mind would invest in solar and wind installations knowing that curtailment is part of the system? It should be possible to give those investors financial support for the curtailment, but then this also has to be added as the cost of a 100% solar and wind grid.
There are also the grid costs in the spreadsheet. It is a meager $279 billion / 39 years = $7.15 billion per year for the entire continental US. Nitche assumes that strengthening of the grid is not necessary, only the cost of expansion of solar and wind is needed.
This seems very low to me. Let me illustrate this again with an example here from Belgium. Just a half year ago, our Minister of Energy approved a raise of the cost of electricity because the grid manager urgently needs to make huge costs in order to accommodate more renewable energy on the grid. This would cost €7 billion over the next decade (we recently learned that this amount will be even higher). This money would be used to build more interconnections with other countries and strengthen the backbone of our grid to bring that renewable electricity inland. So, this doesn’t even include expansion of the grid.
That is somewhat more than €700 million per year for a higher share of renewables on the Belgian grid or roughly about one tenth of the cost that the US supposedly needs to get 100% solar & wind with methane backup.
Is that in proportion to the size and population of both countries? Well, Belgian has currently about 12 million souls compared to 330 million in the US. Belgium has an area of 30,689 km2, compared to 9.8 million km2 of the US. The map below shows the comparison between Belgium (blue) and the US (red). I will leave it to the readers to try locating the blue dot on this map.
This makes me rather skeptical towards the idea that 7.15 billion per year would be enough to reach a 100% renewables US grid. Based on the experience that Belgium is currently going through, I think 7.15 billion is woefully short of what will really be needed.
Concluding: the reason why this calculator will never ever give a realistic cost of a 100% renewables grid is that the assumptions it is based on are not realistic in the first place. The averaging of all numbers erases actual problems like intermittency and seasonal storage. Then it is pretty easy to claim that a 100% renewable electricity grid is not only possible, but also cheap. Using this (flawed) methodology, there is no need for methane storage in the system, because as long as there is enough energy produced on average, everything seems okay. In reality however, storage is necessary and, because of the seasonal differences between production and demand, will not be cheap.
Basically, the model doesn’t represent the working of an actual grid and therefore can’t possibly determine the cost of a 100% renewable electricity grid.
via Trust, yet verify
November 30, 2023 at 05:19PM
Iowa Climate Science Education 