In this second post on the viability of solar + storage (first post here) I examine the conclusions of the Aurora Energy report, which presents a rather more optimistic view of the prospects of the combined technologies, claiming that the cost of backing up solar generation and integrating it into the energy system was negligible, representing just 2% of the current cost of solar energy.
Costs of battery storage are falling…but is it enough?
The Aurora report assumes that large-scale deployment of battery storage will significantly reduce back-up costs of solar in the U.K., suggesting that if costs fall to £100/kWh by the early 2020s, as current trends suggest, then 8 GW of solar capacity would have a negative cost of -£3.7/MWh i.e. be a benefit to the system. This is for two reasons: the existence of batteries on the system acts to smooth out energy and balancing prices, reducing the costs of intermittency, and also provide capacity thereby reducing the need for back-up generation.
“Batteries and solar are a complementary combination. With batteries improving the capture prices of solar, and solar creating a generation profile whereby batteries can profitably store and then deliver to the market as needed.”
In its analysis, Aurora considers the costs of intermittency, back-up generation and balancing costs. In terms of intermittency, this is measured based on the weighted average wholesale power market price in all the hours of a year when power is typically delivered by a given technology, compared with the baseload price. The legitimacy of this approach could be undermined if the wholesale prices used don’t contain the full costs of intermittency, which can be seen through balancing market as well as energy market prices.
The implication of the report is that these costs are excluded as the authors go on to say that failure to deliver in any settlement period is met by asset owners via cash-out prices. The fact that balancing costs are met by asset owners does not justify their exclusion since owners will require returns from their market contacts and/or subsidies to compensate them for these costs, and therefore they are passed back via those mechanisms.
Balancing costs also go beyond cost of non-delivery in a specific delivery period. It is also necessary to adjust for changing near-term expectations where more cloud cover over a period of hours would require the system operator to pay other generators to come online at prices that can be significantly higher than energy prices seen in the within-day wholesale market. It is not clear how these costs are incorporated by the analysis.
Aurora uses the cost of generation provided through the capacity market as a means of determining the cost of back-up generation. This could again be problematic depending on the assumptions used to model the capacity auction outcomes as the market rules change. To date the auctions have out-turned too low to incentivise large-scale new gas generation onto the system, the absence of which leads to some concerns about future capacity margins.
The final area of cost considered by Aurora relates to balancing costs. As noted above, the direct impact of cash-out costs on solar generators is excluded, however, divergence between delivery commitments and actual delivery impose additional costs on the entire system as well as the immediate imbalance costs faced by the responsible generator, and these are considered by the report. However, the overall finding was that the sources of back-up generation required by intermittent solar delivered such additional flexibility that overall balancing costs are reduced, hence this segment is treated as offsetting the other two sources of cost.
By not including the costs of subsidies in its analysis (at least they are not discussed in the report), the Aurora approach risks understating the costs associated with provision of a balanced low carbon system for the reasons outlined above. The analysis also appears to exclude the costs incurred by the system operator in managing balancing and frequency, which should not be excluded when considering the real costs of maintaining security of supply in a system with significant solar generation.
The amount of storage required is likely to be large
The two reports described in these posts take different approaches to the same question – can the intermittency issues of solar power be addressed through use of electricity storage. These different approaches make it somewhat difficult to compare their arguments – Adam Smith focusing essentially on security of supply while Aurora focuses on cost. The Adam Smith report suggests 10 GW of storage capacity would be needed to back up its model solar fleet, while Aurora suggests 8 GW battery storage could come online by 2030, although it does not specifically say whether that would remove the need for backup generation.
Energy Matters has explored this question, and finds that:
“installing enough battery storage to convert intermittent wind/solar generation into long-term baseload generation increases total capital costs generally by factors of three or more for wind and by factors of ten or more for solar, even at (battery costs of) $100/kWh.”
However, like Aurora, it does not explain its methodology in arriving at its estimates of required storage capacity.
Further analysis worth considering is outlined in this blog from Brave New Climate, which examines the question of storage with renewable energy from the perspective of energy returned on energy invested, or EROEI – in other words, measuring the energy produced by a form of generation versus the energy required in its construction. Drawing on the conclusions of Weißbach et al., Energy 52 (2013) 210, the author illustrates that solar, wind and biomass when combined with storage have EROEIs lower than the minimum hurdle level suggested in the research.
This presentation from Liten, a European research institute driving the development of sustainable energy technologies, outlines further challenges with batteries, focusing on the challenge of producing the amount of batteries that would be needed to support renewable generation, and raising questions about availability of raw materials and the sustainability of the production process.
Finally, this rather long and detailed post from the Energy Collective seeks to examine the economics of battery storage using actual technical data, and concludes that the actual costs of using batteries to back-up solar generation are far higher than the headline numbers, and in a domestic setting would not be efficient unless time of use pricing had wide differentials between day-time and night-time rates:
“Storing a quantity of high-value, on-peak solar energy during the day, to retrieve a smaller quantity of low-value, off-peak energy during the night, is not smart, unless the rate differential and/or subsidies are extremely high.”
So how viable is “solar + storage” really?
When I began researching this post, I felt confident in the prospects for storage as a back-up for renewable generation, although I was aware of the capacity constraints of the domestic offer. I feel now as if I’ve opened a Pandora’s box of conflicting research, analysis and propaganda. Solar + Storage is a beguiling concept, but I feel that much of the hyperbole surrounding it is premature when set against both the current installed base of storage and its technical capabilities.
It is not clear to me that falling costs alone will make batteries the solution to solar intermittency, as the size of battery required to meet actual demand over meaningful periods would also have major space implications. Current domestic batteries as noted above cannot power high consumption appliances for longer than a couple of minutes, and are typically the size of a dishwasher giving some indication of the challenge.
Of course innovation continues in the area (beer batteries, anyone?) so more could be possible in the future, and chemical storage is not the only alternative…thermal storage approaches may prove to be more effective. For now however, solar power will continue to need conventional generation as a backup.