So, what is their LCOE (levelised cost of electricity). Well that depends on how much you can discharge them each time. This very interesting blog post discusses battery packs and longevity. It's a long post but well worth reading. In essence, you do not want to completely discharge or completely charge a lithium-ion battery because battery life is shortened dramatically if you do. You also don't want it to get too cold or too hot. The Tesla car battery system has coolers and heaters to extend battery life so it seems logical that the Powerpack has them too. (This article suggests that battery degradation in Teslas is remarkably low, even after 150,000 miles)
So let's assume a 60% DoD (depth of discharge) for the Powerpacks with software controls that prevent it going above 85% and below say 20% except in an emergency (with manual overrides). This would allow far more cycles. Note that at the end of 10 years, the battery should still have 80% of it original capacity, but I haven't included that in the costings as new batteries are likely to be much cheaper then--old batteries even with 80% of original capacity will be competing against new batteries which will cost 20% of what batteries cost now. On the other hand, they'll still be usable.
Let's say 10 years, or 3650 cycles, DoD 60%. That gives a LCOE of around $230 per MWh. Now according to Lazard's analysis (see next blog post) electricity generated by peaking power gas costs $163 to $218 per MWh. So in other words, batteries (at least, Tesla's) now cost just a little bit more than the top end of gas peaking.
But batteries have several advantages over gas peaking:
- Batteries are 100% green. Tesla's gigafactory will be 100% powered by solar, and Tesla will be recycling the batteries when they get old.
- The response time to fluctuations in the grid is within micro-seconds with batteries, but it can take several minutes for a gas peaking plant to get going.
- The cost of batteries is known. Once installed, there is some minor maintenance but the fuel costs are zero. A gas plant has to pay for its gas, which fluctuates in price. Price certainty is always valuable and to be preferred ceteris paribus to price fluctuations.
- Batteries are modular, and have a surprisingly small footprint. So they can be placed next to the generation plant (e.g. a wind or solar farm) or at a substation. They can be spread across the grid, reducing the need for grid upgrades as demand rises.
- Gas peaking power is useful when demand is "too high" but batteries are useful when demand is both "too low" and "too high". When the wind is strong at night but demand is low, or when in midsummer solar produces "too much" power because baseload power plants can't be switched off, so that the wholesale price plunges or even goes negative, batteries can store that power for when demand is "too high".
Battery costs have fallen 70% over the last 18 months. Before the last 18 months, battery costs were falling by 15% per annum. So, let's be conservative and assume costs go back to a 15% per annum rate of decline. That means that in a year's time, electricity from Tesla's Powerwall will cost $195 per MWh. In two years' time, it'll be down to $165 per MWh--the same as the cheapest gas peaking. In three years, $140 per MWh. Let's do a back-of-envelope costing of a grid powered by solar and wind at $40 per MWh in 3 years' time (costs falling by 10 to 15% per annum), with 8 hours of battery storage. That gives us $40 plus 1/3 of $140, or $87/MWh, not much more expensive than the cheapest coal, even excluding carbon taxes.
Does that mean we can go to 100% renewables without other solutions? No, not quite, though we could easily go to 70%. Batteries will cope with daily demand fluctuations but not seasonal ones. Power to gas (the Sabatier process) might be part of the solution, in which case we may keep our peaking power gas plants. On the other hand we may not: it's much more efficient to produce hydrogen by electrolysis and burn it later when needed that to produce hydrogen and then add the next step of producing methane. Hydro is another: Australia's Snowy River Hydro system for example will most likely end up producing most of its power in winter, to provide for winter electricity demand in SE Australia, not spread it across the year as it does now. CSP (concentrated solar power) yet another.
One last point. The introduction of batteries into the grid will put coal and nuclear at a disadvantage. Not only is the average cost of wind and solar cheap, the marginal cost of wind, solar and batteries is virtually zero. On the other hand, because coal and nuclear can't easily be scaled up or down, and because there is fuel cost involved, their marginal costs are not zero. So any grid operator/utility at the margin will favour wind and solar if it has battery storage. This will increase the costs of baseload generators like coal and nuclear because their costs will have to spread over fewer hours and smaller capacity. You can see this already happening in China. Batteries will force the grid towards renewables, because renewables are cheaper and preferring them will make baseload power even more expensive. And that process is starting now.