This page is supplementary material to the post “Electricity in Texas part II – the cost of a 100% renewable grid” on Climate etc.
Different battery technologies could be the cheapest option by 2030. One or more lithium ion variants are assumed to be in contention. This section uses current information as a base for future costs. Many reports on battery storage costs produced in the last two years have been overtaken by recent commercial cost disclosures and at some point they information here will also be overtaken by events!
For grid use, to preserve the battery cell lifetime, the DoD (depth of discharge) may be restricted to 80%. This applies to lithium ion but not necessarily to other technologies such as sodium sulphur, sodium ion or lithium sulphur. The charge and discharge rates for grid use will already be low – occurring over at least 5 hours or 0.2 C where 1.0 C is defined as total charge or discharge over one hour. The grid hourly simulation defaults for the 100% renewables scenario show tier 1 battery storage cycling in 2.6 days on average.
The estimates for current and 2030-2040 prices below are split into DC battery pack, inverters , charging/control/environment and maintenance.
DC battery pack current lowest cost – $190 / kWh, 2020-22 – $124 / kWh
At a conference earlier this year General Motors let slip that LG Chem would supply lithium-ion battery cells at a contract price of $145 / kWh in 2016 for the Chevy Bolt. They also said that the cost of battery cells would be $100 / kWh by 2022. $145 / kWh was half the price previously estimated by outside analysts who have now been forced to recalculate the Bolt profitability.
Battery cells are only one part of battery packs. Battery pack costs are typically 20-40% more than raw cell prices, giving a mid-point of $190 / kWh.
Secondly the head of Tesla’s investor relations group dialled in to a conference and said that the Tesla model 3 (now being delivered) will have a battery pack cost of less than $190 / kWh. This was after an analyst on the same conference call claimed that the Tesla cost was $260 / kWh.
At a later event at Tesla’s new Gigafactory in Nevada, Elon Musk claimed that Tesla and partner Panasonic were on a path to achieve $100 / kWh battery pack prices by 2020 – far earlier than the rest of the industry, and five years ahead of industry predictions. Some commentators interpreted the words (which do not seem to be on record) differently, as meaning the cell (not pack) prices will be $100 / kWh, giving 2020 pack prices around $124 – before the year 2022 in which the Bolt battery pack could hit a similar cost. Tesla has also verbally amended a statement on the battery price reduction expected with 2020 completion of its Nevada gigafactory up from 30% to 35%, which would give a battery pack price of $124 / kWh which is the basis for the current DC battery pack estimate.
Another company, EOS is now offering forward pricing for 2022 of DC battery packs for grid use at $100 / usable kWh. This includes battery modules, the battery management system and outdoor enclosures. The charging and inverters for full grid integration are extra (see costing for these components below).
Current re-use of electric vehicle battery packs – $100 / kWh
A company, Freewire, is already buying second-hand Nissan Leaf batteries for $100 / kWh. Except when being fast charged, EV lithium-ion batteries are typically not highly stressed during normal driving with home charging, and most packs will have a thousand or more cycles of additional lifetime at 80% DoD when reused as grid storage.
V2G (Vehicle to Grid)
Some battery storage might be provided by grid-connected electric vehicles whose owners consent to the battery being used to fill gaps in renewable demand. The owner might specify unusual travel plans and the smart grid and car would ensure the car is sufficiently charged for either regular or unusual journeys, but otherwise the battery would be under the control of the grid. Nissan is suggesting the grid should simply provide free travel charging to car owners who participate in V2G.
This option would reduce the cost to the grid of tier 1 battery storage. Drivers are willing to pay larger sums for incrementally larger / faster / better cars, whereas they would quite rightly complain if electricity bills rose by similar amounts outside their control!
In the cost model no reduction in storage LCOE has been made for a proportion of V2G storage. But the potential is there.
DC battery pack 2030-2040 high LCOE estimate – $108 / used kWh, $9.6 / MWh contribution
Tesla is expecting a DC battery pack price of $124 / kWh by 2020-22 based on a 35% reduction for the Gigafactory volume increase. There is also likely to be at least a further 30% reduction as electric vehicle volumes ramp up in the 2020s giving a high figure of $86 / kWh in the 2030s.
However, to preserve battery lifetime an 80% depth of discharge is assumed, assuming lithium ion variants are used. Thus the effective price per kWh of storage will be $86/0.8 = $108 / kWh, which is the high capital figure for DC battery packs used in the cost model. It excludes inverters, charging, environment and maintenance which are costed separately below.
With 300 GWh of battery storage capacity and a power capacity of 60 GW the charge and discharge rates will be below 0.2C (i.e. minimum charge and discharge times of 5 hours. Factoring in the 80% maximum DOD (depth of discharge) means that the batteries will be even better looked after. With some improvements to the battery technology to improve lifetimes it is highly likely that the batteries will last for at least 2,000 cycles of 80% DOD, equivalent to 1,600 full cycles (100% DOD). In the 100% renewables grid hourly simulation the batteries are cycled once every 2.6 days on average, so would last some 2,000 x 2.6 = 5,200 days or 15 years. Based on a 6% average cost of capital the discounted lifetime is 9.7 years, giving a contribution to the average LCOE of $9.6 per MWh supplied by the grid.
The 2017 BNEF (Bloomberg New Energy Finance) Summit presentation of Michael Liebreich, slide 54 gives a 2030 EV battery pack cost of $73 / kWh. Converting to maximum 80% depth of discharge gives a $91 / kWh battery pack price for 2030. The BNEF figure is based on a 2016 battery pack cost of around $280 / kWh (significantly above the current $190 / kWh figure from two separate industry sources used above), but with a 19% learning rate reduction (higher than used above). This is taken as confirmation that $108 / kWh is a conservative estimate as intended.
DC battery pack 2030-2040 low LCOE estimate – $40 / kWh, $4.7 / MWh contribution
In the 2030-2040 time frame it is highly likely that second-user battery prices will be in the range of $40-60 / usable kWh (down from $100 / kWh currently on a very small market as mentioned above).
To make an allowance for the wear caused by the initial use in EV’s, the expected lifetime has been reduced from 15 years to 10 years. The discounted lifetime is then 7.4 years and the contribution to LCOE becomes $4.71 / MWh.
Some believe repurposed EV batteries may find it difficult to compete with new batteries. For this to be true, new battery storage would have to result in a similar LCOE contribution – probably being a little more expensive with a longer lifetime than repurposed EV storage.
Inverters – currently $120 / kW
There is little battery 3-phase inverter pricing information on the web, so costs for solar PV 3-phase inverters were used instead. Fortunately the LCOE contribution will be small so the accuracy doesn’t matter too much. Three phase utility solar inverters (table 2 page 9) currently cost $120 / kW.
Inverters –2030-40 high LCOE estimate $100 / kW, $1.5 / MWh
Assumes a conservative reduction of 17% from today’s 3-phase utility solar PV inverter cost.
Inverters –2030-40 low LCOE estimate $60 / kW, $0.9 / MWh
Assumes a 40% reduction compared with the high cost i.e. a 50% reduction compared with current 3-phase utility solar inverters.
DC charging, control and environment $30, $50 / kW, LCOE $0.5, $0.8 / MWh
Assumes 50% of the inverter prices for low and high LCOE. Again the LCOE contribution should be low so the assumptions and accuracy are not critical to overall pricing.
Battery storage maintenance LCOE $0.8, $1.8 / MWh
Assumes 15% per decade of the total battery storage costs (DC battery pack, inverters,+ DC charging + control + environment) for both the low and high cases.
Savings from Tier 1 battery storage
There are two sets of savings from the tier 1 battery storage. One set involves higher efficiency and lower thermal stress due to the lower gas turbine generation ramp rates and reduced number of starts. This has not been quantified.
The bigger savings are from a reduction in back-up gas generation capacity.
Reduction of back-up gas turbine capacity
Tier 1 battery storage allows back-up generation to be only a little larger than the average over 24 hours of the largest capacity deficit after subtracting wind and solar generation. This arises because back-up generation can be started before required and the surplus stored in tier 1 battery storage to reduce the required capacity of the peak deficit. Some of the peak deficit would come from batteries. Backup capacity does not need to meet the maximum hourly peak deficit from demand less renewable generation.
The difference could represent the cost of around 15 GW of gas generation back-up capacity., estimated to save approximately $6 / MWh across all supply. This is nearly half the high tier 1 battery storage total LCOE and nearly equal to the low LCOE – the tier 1 battery storage nearly pays for itself at the low LCOE contribution!
Perspective on tier 1 battery storage costs
2030-40 tier 1 battery storage is cost-effective for some hours of storage (7.5 hours in the 100% renewables solution) of the average load of 40GW, but is too expensive for days or weeks of storage. So there is a need for the cheaper tier 2 renewable gas storage.
Location of battery storage
If CREZ transmission network (80 GW) were really to be over-configured compared to the peak demand (71 GW), battery storage would be most effective in the CREZ regions rather than in the population centres. That’s the conservative base assumption for the purposes of this article. In reality there may be transmission cost savings from a proportion of battery storage at the population centres in the east because the network need not then be sized for peak loads. This would require separate simulation which has not been done. Neither have any transmission network cost reductions been factored in for this.