I'm doing my master thesis about the renewable energies. So my goal is to calculate how much it will cost to switch our fuel based energy generation to renewable energy generation. So my first thought was I use the levelized cost of electricity. I found some forecasts of how they will develop to 2050. So I made different scenarios in my thesis on of fast/slow the switch will be and now wanted to calculate the cost of them. So what I wanted to do is take those levelized costs of electricity (LCOE) and multiply it with the amount of electricity production different technologies have in my scenarios. After I did all those calculations I started to become sceptical if this all makes sense at all. Because I thought to myself if I look at the formula of LCOE, if you would change the amount of generated electricity it will also effect your LCOE, so lets say I produce more electricity through coal, this will decrease my LCOE and so on.

I am kinda stuck in the middle of nowhere now, because one part in me tells me that this are basically the costs per unit so it makes sense what I did and the other thoughts I have is, it makes no sense.

in other words: Can I use the LCOE forecast of 2030 from those studies and multiply that with the amount of electricity produced for example in my 'scenario B' in the year 2030 ?

sidenote: scenario A, B, C and D have a different approach of how fast/slow the switch should be done, so lets say in scenario B I decided that in 2030 there will be already 60% of renewable energy, while scenario C is more conservative and the renewable energy is still beyond 20% of the whole generation.

Can someone please help me with my logical dilemma ?


In general, an LCoE forecast for a year, is the unit cost of electricity for generators that come online in that particular year.

Whereas, for a given year's (e.g. 2030) forecast of the cost of electricity, the average unit cost of electricity is a volume-weighted arithmetic average of the levelised costs of all the generators that have been built previously, that are operating in the forecast year. Typically, those generators will have been built in the previous 20 years.

That's not the only challenge:

  • you will need to check the actual definition of the LCoE data that you have. It might be the unit cost for generators that get their Final Investment Decision in that year (and so will come online sometime later, probably in the range of 1-15 years)
  • usually (but not always), changes to the underlying cost will affect total demand, so you've got a two-way feedback mechanism between quantity and price.

  • that rate at which things are deployed, and the consistency of speed with which they are deployed, affects their LCoE. There are chokepoints in various supply chains, there's exogenous and endogenous cost reductions from learning by doing, and so on.

  • merely averaging all the individualised costs of energy, omits the costs that are borne by the system itself: the cost of transmission and distribution networks, the costs of storage, and the costs of mechanisms to keep demand and supply in perfect balance every 20 milliseconds (or so) of the year.

So you can take historic LCoE, and multiply it by the quantity of electricity demanded in a year, to calculate total cost of electricity consumed in that year, subject to the above caveats. Note that that will give you a different answer to the amount of capital investment needed.

| improve this answer | |
  • $\begingroup$ Thanks for that really good and detailed answer thats why I gave you already a +1. Still I need that little sugar from you, since you seem to have a good knowledge in this topic. I won't have the capacity to dig into this that deep that I really get close to what is going to be realistic in the end. But you were right actually even without telling that here, those forecast specified two different LCOEs the ones for already operating generators and the one that may be built new in 2030 and so on, I would take the LCOEs of already build generators and multiply that with my elctricity production $\endgroup$ – MaradonaAtCoding Feb 27 '18 at 14:15
  • $\begingroup$ would this make sense just mathematically even though you are right there are deeper two-way feedback relations that will influence that again aswell but I could write that into my thesis as an assumption that I leave those other things out which of course will fake the result a bit, but since I am student on my own and not a whole organization existing out of 6-7 employees who could also take that into account, is that reasonable for you ? $\endgroup$ – MaradonaAtCoding Feb 27 '18 at 14:17
  • $\begingroup$ @MaradonaAtCoding I've added a final paragraph, does that help? $\endgroup$ – 410 gone Feb 27 '18 at 15:16
  • $\begingroup$ thanks a lot makes sense to me and I think I understood it ;) $\endgroup$ – MaradonaAtCoding Feb 27 '18 at 17:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.