This morning I noted (and tweeted) that there was virtually no wind production collectively coming from all wind farms across the mainland.
This situation has continued through till this afternoon, which means that for the past 15 hours total wind production NEM-wide has been below 400MW, and as low as 200MW (with most of what is being produced in this spell coming from Tasmania). What’s more, the total wind production has been below approx 550MW since midnight heading into Saturday morning.
This can be seen in the following snapshot from NEM-Watch v10, highlighting how the current lull (approaching 62 hours) is longer than patterns seen over the prior 21 days:
Note that I have not had time to do any definitive review of the longest period at which wind output has been low (such a review would turn up different durations for different definitions of “low”). For the purpose of these back-of-the-envelope calculations, this is a useful starting point.
I’ve posted previously about the highs and the lows of wind production (including recently in terms of interaction with solar, given the inclusion of solar PV (small and large) in the new NEM-Watch v10) but today I have wondered what this means in energy terms – and the implications for storage.
Let’s assume, for the moment, that it’s only wind that supplies the NEM – a very extreme case that not a single person is advocating (on the left, or the right), but which I have found useful as an exercise to get my head around the size of the storage challenge.
Since 1st November 2014, the daily average NEM-wide Scheduled Demand has oscillated from a low of about 17,700MW (on Christmas day) to a high of 24,100MW (on 22nd January – the day after I posted this) – as shown in this trend from NEM-Review:
Let’s pick 21,000MW as a rough median level, and knock off 500MW in a contrived scenario saying that the current doldrums would occur at a time when demand averaged this level – this means, in energy terms:
20,500MW x 62 hours = 1,271,000MWh of energy required from storage
Note that this assumes all sorts of things, like no alternate sources of anything except for existing wind capacity (whereas we can see in the NEM-Watch snapshot above the effect of hydro, for instance, as another zero-emission source – and one with more flexibility). However it does also assume that the current level of solar injections would persist at the same time, as they are captured (as negative demand) in the scheduled demand numbers in the NEM-Review chart.
The information in this article seems to imply (if my maths is correct) that a vehicle could contribute around 20kWh of storage – and this seems to be in the range of what I have read elsewhere. Leaving aside trickier technical questions (like how quickly injections can ramp up, and how controllable, how they’d be charged in the first place, how many cars could be left in-situ for 2.5 days without use to enable them to feed to the grid, etc…) this means that, if we were to deliver that amount of stored energy from electric vehicle batteries alone, we would need:
1,271,000,000kWh / 20kWh = 63,550,000 electric vehicle batteries required.
How many cars are there in Australia, I wonder? The ABS here notes something like 17,000,000 vehicles in total, or only 26% of the total vehicle number required in this contrived scenario (notwithstanding a smaller number of passenger vehicles, and some not being in the NEM, etc…).
Obviously this does not add up. Now, this scenario is contrived in a number of ways – so I wonder what might make it more “real”?
Continuing from this back-of-the-envelope calculation, we hypothesise that the NEM might be able to absorb about 10 times the current installed capacity of wind – and that this might mean that, at peak wind production, wind farms might supply all the NEM’s consumption at some points in time. Assuming a 10-times multiplier, this might also mean the 500MW minimum output above scales to 5000MW, which might mean 16,000MW “unserved” over the 62 hour period.
16,000MW x 62 hours = 992,000MWh of energy required from storage
992,000MWh / 20kWh = 49,000,000 electric vehicle batteries (still nowhere near plausible).
I’ll return to this thought process as time permits…