In previous post, I detailed my initial struggles to figure out what is happening in the 100% renewable electricity grid calculator. It lead me try a different approach: just focus on the 8¢/kWh result and work myself backwards. That would give me a direct way to figure out the assumptions that were build in the spreadsheet.
The scenario that Nitsche proposes in the calculator is to overdimension solar and wind capacity and to convert the generated excess power into gas that can be burnt in gas-powered power plants to fill in when solar and wind electricity production is insufficient.
The formula behind the end result of 8¢/kWh is pretty straight forward. It sums the total cost over 39 years of:
- Solar capacity
- Wind capacity
- Battery capacity (optional, the proposed scenario doesn’t assign a battery capacity)
- Backup power plant capacity
- Production of fuel via Power-to-Gas
- Grid expansion
It then divides this sum by the total demand over that same period to get to the end result in cent/kWh.
In this post, I will detail how the needed capacity of those backup power plants (point 4) is calculated in the spreadsheet.
That specific calculation is done in cell O18 in the “calculator” sheet:
cost in dollars per MWday × 14,245 × largest deficit ÷ 1,000,000,000
The 14,245 constant is the number of days in 39 years (365.25 × 39 = 14,244.75 ≈ 14,245) and the 1,000,000,000 is a conversion from dollars to billion dollars. So, the cost of backup power plants is calculated as the cost of those backup plants with a maximum capacity equal to the largest gap left by solar and wind.
This makes sense. The spreadsheet calculates the cost of 100% renewable electricity, meaning that the backup capacity needs to be at least large enough to fill in the largest gap left by solar and wind. However, he lost me with the way this capacity is calculated. It is calculated in cell M28 as the largest gap from … January 1 until December 30, 1980…
?!?!?!
These are just the first 365 days of the 39 years period (December 31 is probably excluded because 1980 is a leap year). The data spans 39 years, why not just use the entire period to calculate the needed backup capacity? If larger gaps occur in the following 38 years, then the result will not be a 100% renewable electricity grid anymore.
When the entire period of 39 years is considered, then the largest gap is 402,130 MW (compared to 318,577 MW for the first 365 days of the period), therefor the backup capacity should be 83,553 MW larger than the value used in the spreadsheet. Meaning that the capacity of the backup plants is underdimensioned: the calculated backup capacity would be insufficient to fill in all cases where solar and wind production is lacking (139 times to be exact). It doesn’t make that much of a difference in the cost though (from 8.1 to 8.4 cent/kWh).
But then, why does the spreadsheet only accounts for the first 365 days of the dataset to determine the necessary backup capacity? In the article in which the spreadsheet is presented, Nitsche made this puzzling claim about the demand data in the spreadsheet (my emphasis):
it is just one year’s demand data, copied and pasted 36 times.
Maybe he (wrongly) assumed that only one year needs to be taken into account when determining the largest gap of a dataset with 1 year of repeating demand data?
Whatever the case, in reality the spreadsheet uses 4 years of repeating demand data (not just 1) and Nitsche mentions everywhere else in the article that the dataset contains 39 years of data (not just 36). It is therefor my impression that this spreadsheet is an adaptation of another, similar spreadsheet. Maybe from this study? It not only has 36 years of data with 1 year of repeating demand data, also three of the four co-authors of this paper are co-authors of the Dowling et al 2020 paper.
This is not the only time that I had this impression. There are other indications that this spreadsheet is an adaptation of another spreadsheet.
Next post will be about the production cost of the gas that will be burnt in the backup power plants (point 5).
via Trust, yet verify
September 26, 2023 at 03:34PM
Iowa Climate Science Education 