How much storage—II

(Source)

The CSIRO and Energy Networks Australia have done some detailed modelling to estimate just how much storage will be needed as the percentage of renewables in the grid increases.  Although the analysis is specific to Australia, the general conclusions are applicable everywhere.

The paper analyses requirements by state, because although there are interconnectors between states, the separate state grids are not well integrated into a national grid.  For example, there’s only one interconnector between Queensland and New South Wales, one between NSW and Victoria, and two (one low capacity) between Victoria and South Australia.

Some conclusions:

• Solar is much more predictable than wind, but wind is handy for night generation, so the estimates are predicated on a mixture of wind and solar.  But in the southern states, the swing between winter and summer means that solar has a greater seasonal variation than in northern states, which means more storage is needed, especially at high penetration of renewables. 
• At low levels of renewables penetration, very little storage is required.  This is because there is already redundancy in the grid to cater for swings in demand as well as potential failures in existing baseload supply.
• At 50%, something like 2-3 hours of storage is needed.  By 70% it’s something like 5 hours.  
• The big surge in how much storage is needed is when we exceed 90% renewables generation.  In the northern states, we will need 5-7 hours of storage.  But in South Australia (where 95% of the demand is in the south, even though the geographical area of the state stretches quite far north), and in Victoria, we will need more than 24 hours’ worth.  (Tasmania has lots of hydro, so it needs little additional storage.)
• That’s with existing interconnectors.  If additional power lines connecting the southern states were built deficit areas would “borrow” power from surplus areas, and pay it back when their situations reversed.  The storage required in the southern states would drop to something much closer to what is needed in more northerly states.  It can be raining in Adelaide while it’s sunny in NSW, and vice versa, and windy in SA while it’s calm in Victoria.
• Biomass or gas peaking will be more economical than storage if the low output from renewables lasts more than 8 hours.
• Rooftop solar and behind-the-meter storage will be important in the mix.   By 2030 in Queensland for example, there will be more rooftop solar capacity than there is currently coal capacity.  And it is likely that by then everybody with rooftop solar will also have a battery.  My rough and ready calcs suggest that behind-the-meter batteries will provide at least 3 hours’ worth of storage.
• They briefly mention concentrated solar power, but don't include it in their forecasts or analysis.  However, it seems almost certain that there will be at least one CSP plant in SA in the north of the state in the next couple of years, and if that's a success, there'll be others in other states.  CSP is relatively cheap and provides dispatchable power which fits perfectly with a high percentage of renewables.

[Read more here and here]

I've made most of these points before, but it's nice to see the experts make them too.