Energy storage getting real
Now that renewable energy sources like solar and PV are cheaper than new coal-fired power stations in most jurisdictions (anywhere with either favorable conditions or a reasonable carbon price), the big remaining question is that of supply variability/intermittency. As I’ve argued before, this problem is greatly overstated by critics of renewables who assume that the constant 24/7 supply characteristic of coal is the ideal. In fact, this constant supply produces a mismatch with variable demand and current pricing structures are set up to deal with this. A system dominated by renewables would have different kinds of mismatch and require different pricing structures.
That said, for a system dominated by solar PV, meeting demand in the late afternoon and evening will clearly depend on a capacity to store energy in some form or another. There are lots of options, but it makes sense to look first at relatively mature technologies like lithium and lead-acid batteries. Renewable News is reporting a project in Vermont, which integrates solar PV and storage.
The 2.5-MW Stafford Hill solar project is being developed in conjunction with Dynapower and GroSolar and includes 4 MW of battery storage, both lithium ion and lead acid, to integrate the solar generation into the local grid, and to provide resilient power in case of a grid outage.
The project cost is stated at $10 million, or $4m/Mw of generation capacity.
Assuming this number is correct, let’s make some simplifying assumptions to get a rough idea of the cost of electricity and the workability of storage. If we cost capital and depreciation at 10 per cent, assume 1600 hours of full output per year and, ignoring operating costs, the cost of electricity is 25c/KwH. There would presumably be some distribution costs, given the need to connect to the grid. Still, given that Vermont consumers are currently paying 18c/Kwh, this doesn’t look too bad. A carbon tax at $75/tonne would make up the difference.
How would the storage work? I’m starting from scratch here, so I’ll be interested in suggestions and corrections. I assume that the storage is ample to deal with short-term (minute to minute or hour to hour) fluctuations, which are more of a problem for wind.
How about on a daily basis? It seems to me that the critical thing to look at is the point in the afternoon/evening at which consumption exceeds generation (As I mentioned, prices matter a lot here). This is the point at which we would like the batteries to be fully charged. The output assumption suggests an average of about 12 MWh generated per day. If we simplify by assuming that the cutoff time is 6pm and that output drops to zero after that, the system requires that 8MWh be used during the day and 4MWh at night. That wouldn’t match current demand patterns, but if you added in some grid connected power (say, from wind, which tends to blow more at night) and shifted the pricing peak to match the demand peak, it would probably be feasible.
As regards seasonal variability, this would be a problem in Vermont, where (I assume) the seasonal demand peak is in winter. But in places like Queensland, with a strong summer peak, a system with lots of solar power should do a good job in this respect.
What remains is the possibility of a long run of cloudy days, during which solar panels produce 50 per cent or less of their rated output. Dealing with such periods will require a combination of pricing (such periods can be predicted in advance, so it’s just a matter of passing the price signals on to consumers), load-shedding for industrial customers and dispatchable reserve sources (hydro being the most appealing candidate, given that potential energy can be stored for long periods, and turned on and off as needed).
To sum up, we aren’t quite at the point where PV+storage is a complete solution, but we’re not far off.