Energy storage systems can maximize their value by providing multiple services within a specified timeframe and ‘stacking’ the resulting revenue streams. This is called revenue stacking (alternative names: value stacking or benefit stacking) and has three major benefits that can help making energy storage projects profitable:
Utilization: Most applications do not require continuous availability or operation (e.g. frequency response) and some are called rarely (e.g. black start). Serving multiple applications can therefore enable higher remunerated battery utilization.
Optimization: Remuneration for most applications is volatile (e.g. auctions for frequency response, volatile wholesale prices for energy arbitrage). Serving multiple applications enables to optimize between them for highest revenues.
Resilience: The specific applications that energy storage systems serve may be subject to regulatory changes. Avoiding dependence on a single revenue source can partially protect against the risk of these changes.
Figure 1 provides a conceptual depiction of revenue stacking. In this example application 1 (app 1) is not sufficient to cover lifetime cost. Rather, the stacking of applications 1, 2 and 3 ensures that lifetime revenues exceed lifetime cost.
Figure 1 – Conceptual depiction of revenue stacking. This schematic assumes that lifetime revenues of application 1 are reduced by also serving application 2 or 3; however, revenue stacking is also possible without reducing the lifetime revenues of the original application, depending on the revenue-maximizing operational schedule.
The implementation of revenue stacking in practice is more complex because energy storage systems can serve multiple applications in various ways. Figure 2 to Figure 5 depict the four main archetypes of revenue stacking, including description, real-world examples from the Great Britain power market, key considerations, and relevance. These archetypes can also be combined if technically feasible and allowed by regulators.
The archetype descriptions consider two key technical parameters:
Active power: The provision of active power, i.e. charging/discharging the storage system (stacked bars, primary y-axis)
State-of-charge: The amount of energy stored in the system at any moment in time relative to full capacity (dashed line, secondary y-axis).
Figure 2 – Revenue stacking through parallel provision of multiple applications. Exemplary schematic for 20 MW electricity storage system.
Description: Power capacity is separated into individual parts. These are provided in parallel to serve different applications. In the above schematic, the 20 MW electricity storage system provides 10 MW to application 1 and 2 each in period 2. As a result, state-of-charge (SoC) reduces from 100% to 0%. The system charges in period 3 to recover SoC from 0% to 100%.
Examples: Electricity storage operators in the UK can provide part of their available power capacity for frequency response services like Dynamic Regulation and bid the remaining capacity into the Balancing Mechanism.
Considerations: Operators should not exceed the power committed to the individual applications in order not to compromise any other application contracted for simultaneously.
Relevance: Parallel stacking is among the most common types of revenue stacking. There are numerous examples for various applications in various geographies.
Figure 3 – Revenue stacking through sequential provision of multiple applications. Exemplary schematic for 20 MW electricity storage system.
Description: Power capacity is provided to different applications in different time periods. In the above schematic, the 20 MW electricity storage system provides +20 MW to application 1 in period 2, then charges in period 3 to replenish SoC and then again provides +20 MW to application 2 (depending on the application, the storage system could be rewarded for providing +40 MW in application 2 as its operating point or baseline was at -20 MW just before activation).
Examples: In 2016, UK’s National Grid procured 200 MW of a service called Enhanced Frequency Response. Selected operators did not bid for the entire year but excluded ~300 hours. In these periods they reduced network electricity consumption for large consumers instead, to reduce their demand charges as part of the Triad scheme.
Considerations: Operators must manage SoC appropriately to ensure that provision of one application does not affect provision of the other.
Relevance: Sequential stacking is among the most common types of revenue stacking. There are numerous examples for various applications in various geographies.
Figure 4 – Revenue stacking through sequential provision of multiple applications with one application in opposite direction. Exemplary schematic for 20 MW electricity storage system.
Description: Power is provided to different applications in different time periods. At least one application is in the opposite direction, so that SoC levels can be managed while being remunerated. In the schematic, the 20 MW storage system provides +20 MW to application 1 in period 2, then -20 MW to application 3 in period 3, which also replenishes SoC. In period 4, +20 MW are provided to application 2 (again, depending on the application, the system could be rewarded for providing -40 MW and +40 MW to applications 3 and 2 respectively based on its baseline operating points just before applications 3 and 2).
Examples: In 2021, UK’s National Grid enabled the possibility to sequentially stack contracted frequency response services like Dynamic Containment with bids to the Balancing Mechanism (BM). The purpose of the BM bids is to manage SoC levels after providing frequency response. The opposite approach is even more common: Offer frequency services to charge the system and then discharge into the BM or intraday market.
Considerations: It can be difficult to find two applications with opposite directions that fully enable managing SoC. Often, required power capacity and discharge duration vary strongly between applications. A starting point to identify opportunities are negatively correlated applications.
Relevance: Opportunities are less common than parallel or sequential stacking.
Figure 5 – Revenue stacking through parallel provision of multiple applications with the same power capacity. Exemplary schematic for 20 MW electricity storage system.
Description: The same power capacity is provided to multiple applications at the same time. In the above schematic, the electricity storage system is serving its 20 MW power capacity to application 1 and to application 2 in period 2. It is then charging in period 3 to replenish its SoC.
Examples: Contracts in the Capacity Market and Firm Frequency Response (FFR) can remunerate the same MW of an electricity storage system in the same time period. During a Capacity Market stress event any power capacity provided to selected ancillary services like FFR are deducted from the power capacity required in the Capacity Market stress event without penalties.
Considerations: The ability to fulfil requirements of two applications simultaneously may be limited and there are high penalties if contractual obligations are broken. A good starting point to identify opportunities is to look for applications that are positively correlated.
Relevance: Least relevant archetype, because opportunities are rare and regulations often inhibit providing for alternative applications when power capacity is already contracted.
For simplicity, all four archetypes refer to the actual provision of active power. However, applications can differentiate between payments for the availability of power capacity only and the actual active provision of power. Figure 6 adds contracted availability of power capacity for application 1 (light blue) to the archetype parallel stacking (see Figure 2).
Figure 6 – Revenue stacking through parallel provision of multiple applications. Light blue reflects availability for application 1. Dark blue and dark orange reflect active power provision to application 1 and 2 respectively. An example for application 1 could be frequency response. An example for application 2 could be wholesale market arbitrage.
Application 1 could be a symmetric application, which pays for positive as well as negative power provision. It also pays for the mere availability of power capacity (light blue) and the actual provision of power (dark blue). In this scenario, the 20 MW system is paid for the availability of ±10 MW in periods 1, 2 and 4. It is also paid for providing active power of +10 MW to application 1 and application 2 in period 2. In period 3, the system recharges. This means the system only contracted applications with contractual obligations that can always be met.
Alternatively, a storage system could indicate its availability to multiple applications and receive respective payments. This is shown in Figure 7, which adds the paid availability of power capacity for application 1 and 2 to the archetype sequential stacking (see Figure 3).
Figure 7 – Revenue stacking through simultaneous reservation of power capacity availability and sequential active provision of power capacity to two applications. Light blue and orange reflect availability for applications 1 and 2 respectively. Dark blue and dark orange reflect active power provision to application 1 and 2 respectively. An example for application 1 could be frequency response. An example for application 2 could be frequency regulation.
The storage system contracts its power capacity availability simultaneously to application 1 and 2. This reflects the archetype overlapped stacking for power capacity availability. As a result, the system gets paid for reserving its power capacity to applications 1 and 2 in periods 1, 2 and 4. In period 2 it is also paid for actively providing power to application 1. It does not need to be available for application 1 in period 3 and is only paid reserving its power capacity for application 2, while it charges. In period 4 it then provides active power to application 4.
In this particular scenario, the system can meet its contractual obligations. This is a lucky coincidence, however. If application 2 were called at +20 MW in period 3, the storage system could not provide it (SoC was at 0). Similarly, if application 1 were called again in period 4 with +20 MW (simultaneously with application 2), the system would have not have enough energy to provide for both applications. Therefore, this type of revenue stacking should not be called overlapped, but rather ‘overbooked stacking’.
For applications where actual activation of active power is subject to a tendering process ‘overbooked stacking’ may be manageable, i.e. high-priced bids would mean the storage system is not called into activation although being paid for availability. However, in most cases this type of revenue stacking would likely be illegal and result in payment of penalties. Also, the high level of transparency in system-relevant applications through obligatory real-time communication of system operation parameters may mean that customers will uncover overbooked stacking quickly.