Sunday, September 4, 2016

More on electricity costs

The go-to people for electricity costings are Lazard, one of the premier investment banks in the world.  They have been publishing their annual "Levelized Cost of Energy Comparison" for 9 years.  They do this for their clients, but after 9 months, as a public service they make these estimates available on their website.

Here is a screenshot of their latest chart (November 2015), with three of my additions (click on chart to see a larger more readable version)


The magenta dot shows the latest PV auction results for Dubai ($29.9/MWh) and Chile ($29.1/MWh).  Two things to note with this. Just 18 months ago ago, a similar bid produced a price of  $59.8/MWh in Dubai.  Lazard's own estimate for 2017 solar LCOE (the grey diamond to the right of the magenta circle) was $46/MWh (Please note that I had to move the circle a little to the left so as not to obscure the '$46')  Of course both Chile and Dubai are in places with higher solar radiance, as the chart below shows.  But the critical point is that costs have HALVED over the last 18 months in these places: the solar radiance on the other hand hasn't changed.  My guess would be that the places on the map with light green solar radiance would have solar which would cost twice that of places like the Atacama Desert and Dubai.  And that's still cheaper than coal and about the same as gas (gas in the US is much cheaper than gas elsewhere)

(Source)


The green dot in the "Solar Thermal Tower with Storage" (i.e., Concentrated Solar Power or CSP) represents the planned CSP plant in Dubai at "under 8 cents per kWh".   Because it can store heat in the form of molten salts, CSP can deliver power 24/7: better, it can deliver power on demand so that when demand on the grid is high or supply low it can act as a peaking power plant.   The cost of the first large-scale CSP plant at Crescent Dunes in Nevada was $135/MWh.   This is in the middle of the band of $119 to $181 given by Lazard.  The South African CSP plant is costed at around $120/MWh, Chile is estimated to be below $100/MWh.  The Dubai CSP plant will produce electricity at below $80/MWh.  Note that the stored molten salt "can remain hot for months".  Heat loss from the stored molten salts is just 1 F per day.  Wow.

The green circle to the right of Lazard's gas-peaking estimates represents my guesses at the LCOE of the Tesla Powerpack.

To sum up:

  • After energy saving, wind is the cheapest
  • Industrial scale solar is next, and the costs there have fallen precipitously over the last 12 to 18 months
  • CSP has fallen 40% in costs over the last couple of years
  • Batteries are now nearly competitive with gas peaking power, and anyway batteries have several advantages over peaking power gas
  • Gas combined cycle is now less competitive than it was just 2 years ago.
We keep on coming back to the same conclusion: there are no technological or economic constraints to achieving 100% renewables electricity generation.  True, the sun shines less at high latitudes, but the wind blows more.  So let's take a "typical" grid.  It would be 1/3 solar PV, 1/3 wind and 1/3 CSP, ignoring for the moment hydro, existing nuclear, etc.  That would cost in the US (leaving out the tax credits for solar and wind) .33*$50 (wind) + .33*$40 (solar) +.33*$140 (CSP) = $77/MWh,  a little bit more than  existing electricity price to industry in the US ($70) .   And I haven't chosen the cheapest costs but the midpoint between the (new) lowest and the highest in Lazard's costings above.  And it also ignores batteries, which are already being used as a cheaper alternative to grid upgrades and as a direct replacement for poles and wires, but is now also being used as a spinning reserve.  Moreover, these costs are falling every year, inexorably.  Renewables are already highly competitive with fossil fuels.  And they will just go on getting cheaper.

2 comments:

  1. Interesting figures, but I note the Lazard chart says "... does not take into account ... reliability-related considerations (e.g., transmission and back-up generation costs associated with certain Alternative Energy technologies)".

    This must surely make intermittent & highly distributed technologies like wind, that have relatively higher back-up and transmission costs, look better than they really are. This strikes me as an odd omission if they are trying to compare different technologies and possibly invalidates the comparisons.

    I understand that these costs are difficult to quantify, but surely they are very important here. If Lazard are a respected source for this information, what am I missing?

    ReplyDelete
    Replies
    1. The problem with costing "transmission and back-up generation " (and storage) for Lazard is that they don't know which one is going to be used. Will it be hydro, CSP, batteries, peaking power gas, or a more integrated grid? Each has different costs.

      This article ( https://theconversation.com/the-cheapest-way-to-scale-up-wind-and-solar-energy-high-tech-power-lines-53597 ) suggests that installing a network of HVDC lines across America would actually reduce costs despite the construction costs, while allowing a much higher percentage of renewables than exists now. Most interesting.

      CSP because of its very cheap integral storage will compensate for variability of both wind and solar, which is why I estimates a grid based on 1/3 wind, 1/3 solar and 1/3 CSP. Even without battery storage (still a bit pricey) the average for the three comes in at below the costs for coal.

      The solution to the variability of wind and solar will be larger and more integrated grids, CSP, and some battery storage for grid stabilisation. How much of each will depend on where you are. "Island Grids" (Western Australia, Hawaii") will require much more storage than continent wide grids.

      Delete