Pricing
18 Nov 2021
Imrith Sangha

Benchmarking battery revenues

The revenue opportunities for battery energy storage systems (BESS) are becoming more complex all the time. This is evident in the recent (and upcoming) changes to ancillary services and the increased participation of BESS in merchant markets. While the energy throughput required to participate in ancillary services is low, longer-duration BESS assets are well placed to take advantage of merchant market opportunities. The diversification of BESS revenue streams has opened a discussion - not least in the Modo office - around the best way to benchmark revenue performance. How should we compare battery assets of different durations?

In this article, we look at the different ways we can benchmark BESS revenues, to make comparisons between assets of different energy capacities. We refer to ‘duration’ as the amount of time (in hours) that assets can operate at their rated power, at full charge when first energised. For comparison, we’ve used a sample of ten assets, listed in Table 1 (below). These ten assets have been selected because they are the top-ranked assets - when revenues are normalised by rated power (MW) - with different durations on Modo’s October Leaderboard.

Table 1 - The assets we will be comparing in this article.

The three different methodologies for normalising revenues that we’ve used in this article are:

  • Method 1: Normalising by rated power (£/MW).
  • Method 2: Normalising by energy capacity (£/MWh).
  • Method 3: Normalising by rated power (£/MW), using a capital-expenditure (CapEx) factor.

Method 1: Normalising by rated power (£/MW)

Figure 1 (below) shows the estimated revenues of the assets listed above, normalised by rated power (MW).

Figure 1 - Monthly estimated BESS revenues (£/MW) of the assets listed in Table 1 (according to the October 2021 Modo Leaderboard).
  • Three of the top five ranked assets are sites with a duration longer than one hour (1h).
  • The two longest-duration assets, Battery 10 (2.25h) and Battery 9 (2h) earned 30.2% and 43.7% of their respective income revenues from wholesale trading. The average of the other eight assets was ~14%.
  • At Modo, we estimate wholesale revenues based on trading in half-hourly and hourly power exchange auctions. Having a larger energy capacity allows longer-duration assets to earn more in these markets (than shorter-duration assets).
  • This is because they are able to provide their rated power for longer durations, meaning they can participate in more auction blocks.
  • Normalising revenues by rated power (MW) subsequently fails to take this advantage into account, thus skewing rankings in favour of longer-duration assets.

Figure 2 (below) shows the estimated Dynamic Containment (DC) and wholesale revenues of our chosen assets in October 2021, when normalised by rated power (£/MW).

Figure 2 - Dynamic Containment and wholesale revenues when normalised by rated power (October 2021).
  • In ancillary services, assets are contracted on a £/MW/hour basis. This means rated power is the determining factor in how much an asset can make in DC.
  • In wholesale markets, longer-duration assets are able to capture larger revenues, and this advantage is more pronounced when normalising by rated power (MW).
  • Because of this, there is a much larger spread in asset wholesale revenues (£6,215/MW) compared to DC revenues (£2,374), due to the skewing in favour of long-duration assets in merchant markets.

Method 2: Normalising by energy capacity (£/MWh)

Table 2 (below) shows asset revenues in October 2021, normalised by energy capacity (£/MWh).

Table 2 - Estimated October revenues (£/MWh) normalised by energy capacity.
(Rank change, left column, refers to the change from the ranking normalised by rated power.)
*Refers to aggregated BSuOS, DUoS and Fixed DUoS charges.
  • When normalising revenues by energy capacity, longer-duration assets such as Battery 8 (1.5h) and Battery 9 (2h) drop down the rankings, as seen in Table 2 (above).
  • This is expected, given that DC was the dominant revenue stream for all assets in October. Therefore, when normalising by energy capacity, assets with a >1h duration have lower figures than they do when normalising by rated power (Method 1).

Figure 3 (below), shows the estimated DC and wholesale revenues of our chosen assets in October 2021, when normalised by energy capacity.

Figure 2 - Dynamic Containment and wholesale revenues normalised by energy capacity (October 2021)
  • When compared with normalising by rated power (Method 1), there is a much larger spread in DC revenues than in wholesale revenues. The spread of DC revenues (£7,415/MW) is ~2.5x larger than the spread of wholesale revenues (£2,834/MW) across our chosen assets.
  • This decreased spread in wholesale revenues happens because normalising by energy capacity offsets the inherent advantages that longer-duration assets have in wholesale markets.
  • The opposite happens with DC revenues. This is because having a larger energy capacity is not an advantage in this market. Therefore, there is a much larger spread in these values than there is when normalising by rated power.

Method 3: Normalising by rated power (£/MW), using a CapEx factor

Normalising by energy capacity (Method 2) has limitations. This is because the ratio of CapEx to installed energy capacity is not 1:1. This means that benchmarking revenues using MWh shouldn't be seen as a reflection of the return on investment any more than normalising by MW. For this article, we have modelled a simple multiplication factor, which we have called the CapEx factor. This is Modo’s view of how CapEx increases or decreases with asset duration for sites of the same rated power, as shown in Figure 4 (below).

For example, the CapEx required for a 1MW/2MWh (2h duration) asset is assumed here to be 1.5x that of a 1MW/1MWh (1h duration) asset.

Figure 4 - CapEx factor.

Table 3 (below) shows estimated asset revenues in October 2021, normalised by rated power (£/MW) and using our CapEx factor.

Table 3 - October revenues normalised by rated power and CapEx factor.
(Rank change, left column, refers to the change from the ranking normalised only by rated power.)
*Refers to aggregated BSuOS, DUoS and Fixed DUoS charges.
  • As we saw when normalising by energy capacity (Method 2), assets with >1h duration rank lower than they do when we normalise by rated power alone (Method 1).
  • However, this drop in rankings and revenues is not as severe as when normalising by energy capacity (Method 2).

Conclusion

Normalising by rated power has - up until this point in time - proven the best method for comparing BESS assets, as revenues have been heavily dominated by ancillary service markets. At Modo, we are always working to improve our product. We’ve changed our Leaderboard so that you can now see revenues normalised by MWh, alongside the usual revenues normalised by MW. We constantly review the ways in which we rank batteries on our Leaderboard, and will continue to do so as the markets for BESS evolve.

As ever, we would love to hear your feedback.