788 x16 Powerwall batteries at 5000 USD, or about 63 MUSD, just in batteries + infrastructure and cooling.
Nearly 100 MUSD!
The price is obviously interesting, but it is difficult to deduce in this case a return on storage, or what amounts to the same cost of MWh destocked. Indeed, the role of this battery (126 MWh of storage for 100 MW of power) is not primarily the storage of energy (the theoretical duration of destocking is 1.26 hours, in practice about 1 hour maximum) knowing that a battery is not used until its complete discharge to preserve its longevity, which limits its possibilities of smoothing the demand or the variations of the EnRis).
The main purpose of this battery is to stabilize the frequency of the network by capping the very rapid fluctuations of EnRis, a battery being able to absorb or restore energy to the network in extremely short times. Its role is therefore to achieve a primary ultra-fast power-frequency setting which makes it possible to correct the gaps before they have consequences on the rest of the network (for lovers of regulation theory, the constants of action time batteries are much lower than network reaction time constants). This system works well, it has been extensively tested by EDF R & D on a test loop of about ten MW and is used with success apparently in several (small) isolated networks of the islands, with powers of the same order.
Functionally, the batteries thus undergo loads / discharges at a high and random rate, according to the cumulative fluctuations of consumption + EnRis. Economically, it is not the stored values that are the main interest, but the stabilization service rendered to the network, which can be extremely large amounts if it avoids blackouts as experienced in the network. South Australian between autumn 2016 and winter 2017 (spring and summer periods for Australia). To fix the ideas, the cost for the community of a blackout in France is estimated by RTE to 25 000 Euros / MWh undistributed !!! I do not know what it is for the South-Australian network, of a very different scale (3 GW of installed power approximately, including imports) but this cost being proportional to the standard of living of the country (it corresponds essentially to the loss of GDP due to the general lack of electricity), it should logically be of a comparable order of magnitude per MWh not distributed.
It remains to be verified that the 100 MW of batteries will be sufficient to stabilize a network whose maximum power can reach 3 GW in order of magnitude. Answer by the facts in a few months, but we can think that the managers of this network have made good simulations … It is to wish for them and for … Elon Musk !
One of the constraints of the batteries concerns the permissible discharge current, just like the charging current.
So the amount of energy stored is interesting data and, in this case, 126 MWh is a high value. However to see the ability to maintain the network frequency during sudden changes, it would be good to know the maximum intensity of discharge.
To get an idea, we can compare the capacity of the battery (126 MWh) and power that allows itself to be drawn (100 MW) to those of the battery (the same type – Li-ion) which is mounted on a Tesla Model X P100D, ie 100 kWh resp. 568 kW.
On the car, it is authorized to pump power proportionally much larger, because this power is absorbed during acceleration, necessarily short, much shorter than the periods during which the battery will be 126 MWh solicited at the level of 100 MW.