Introduction

In this article we consider the role and application of battery energy storage systems (BESSs) in supporting renewable energy power generation and transmission systems and some of the challenges posed in seeking to project finance BESS assets.

The need for energy storage

Not so long ago, someone asked the following question at a conference on the development of African power networks attended by one of the authors: why can't we just use renewables to meet Africa's demand for electricity? There is, after all, abundant solar radiation across most of the continent. There are obvious challenges - it is dark at night and the winds do not always blow (and sometimes blow too hard for wind turbines), creating variations in generation capacity and, in deregulated electricity markets, price variation and volatility. BESSs offer a number of attractive solutions for shorter-term energy storage to spread supply capacity over time and to enable electricity price arbitrage.

Batteries are relatively cheap for smaller scale and shorter duration energy storage and prices of cells have historically fallen. They are also well-suited to energy storage in that their “round trip” efficiency is high (around 83 to 86% for conventional lithium ion[1] and up to 93% for lithium iron phosphate (LiFePO₄) batteries[2]), which is slightly better than pumped hydro (70 to 80%) and much better than compressed air systems (42 to 67%) or compressed green hydrogen (18 to 46%, depending on the re-conversion method).

There is also a more obscure technical challenge associated with relying predominantly on renewable power generation: the need for “inertia” to ensure grid frequency stability.

What is inertia?

Ignoring HVDC transmission lines and interconnectors, electricity distribution networks operate using alternating current (AC). AC is used because it is easy to transform between different voltages using a transformer: high voltages are needed for transmission lines to minimise energy losses and lower voltages are required by consumers and other users for safety and practical reasons. In order to avoid causing problems with and possible damage to connected equipment, the frequency (typically 50 or 60 Hz), phase and voltage of the grid must be fixed within narrow tolerances.[3] 

Traditional power generation systems, such as thermal power stations, utilise turbines and generators with large rotating masses which have significant real inertia, storing large amounts of kinetic energy and physically resisting changes in rotational speed. Once a generator is synchronised to the grid, this inherent inertia helps to stabilise the frequency and voltage and to slow down changes in frequency caused by changing electrical loads or supply disruptions. This property is known as “inertial response”.

If thermal generation systems are replaced by renewables such as wind[4] and solar, and the inertia response of the grid is not replaced by other inertial systems, the grid may become more vulnerable to voltage and frequency deviations that exceed permitted limits; and such excursions may trigger disconnections of generating units or other shutdowns.

In particular, certain types of wind turbine generators have a design which disconnects from the grid when the voltage falls below a minimum threshold. The distributed nature of wind turbine generators makes them vulnerable to a “chain reaction” effect which may result in a cascading disconnection of turbines from the grid.

A striking example was the South Australian blackout that occurred in 2016 following extensive storm damage to the state’s electricity transmission network. Almost the entire state lost its electricity supply as successive transmission lines and wind farms and ultimately the high voltage Heywood interconnector to Victoria tripped out owing to cascading voltage and frequency “events”. It took several days fully to restore the electricity supply to the entire state, relying initially on the Heywood interconnector to establish an initial stable state and to restart the Torrens Island Power Station because local black-start facilities were insufficient.

Synthetic inertia

Battery energy storage systems have a very useful property: using appropriate electronic control systems, high-power inverters and step-up transformers to convert their direct current (DC) output to AC at grid voltage, power can be transferred into the grid in a flexible, actively directed manner, that is able to respond dynamically and almost instantaneously to grid deviations in frequency and voltage. Such systems are in effect a form of “synthetic inertia” but offer greater flexibility than traditional “spinning” systems.[5] 

The UK National Energy System Operator is developing a framework to procure a suite of “dynamic response services” from service providers, comprising dynamic containment (DC), dynamic moderation (DM) and dynamic regulation (DR) services which are planned to work together in concert to control grid system frequency and to maintain it within permitted limits, replacing to some extent the traditional inertia provided by thermal power stations. The intention is to create day-ahead frequency response markets for DC, DM and DR.

These new services are expected to be provided by energy storage systems and battery systems are well-suited to perform such roles owing to their fast response times. However, managing the battery state of charge (SoC) in advance and keeping systems within their warranty constraints (see below) poses technical and commercial challenges. Dynamic response products may also need to be “stacked” by providers (with a single BESS providing different services simultaneously, but with each MW of capacity partitioned to provide a single service) to optimise utilisation and revenues, leading to additional complexity.

BESS services more generally

BESS has many potential applications other than dynamic response services which are well suited to commercial exploitation. Notable examples are the following:

• provision of additional electrical supply capacity at times of system peak demand;
• energy time-shifting (allowing arbitrage between higher and lower energy prices);
• transmission system congestion relief (acting as an energy sink to spread the demand on a transmission line over time);
• voltage support;
• black start services to provide the initial power required to start up larger power plants (for which a provider may be paid for availability, even if their services are rarely used);
• transmission/distribution systems upgrade deferral (similar to congestion relief services); and
• demand side services, including power reliability/UPS systems and power quality services.

Economics of BESS services

It is important to keep in mind that in economic terms, most BESS revenues are typically derived from time-shifting/price arbitrage, congestion relief and providing security of supply. Other services, including dynamic response/synthetic inertia typically provide a relatively small component of enduring revenue streams, despite their critical role in ensuring grid stability (not least because the volume of these services is relatively low, and even a small volume of market entry by flexible capacity can reduce the market clearing price for these services).

However, time shifting/price arbitrage, congestion relief and even dynamic response services are likely to involve merchant risk. For example, price arbitrage involves active trading in wholesale markets, and the UK is proposing that dynamic response services would be priced by day ahead auctions.

While operators of BESSs may “stack” different merchant revenue streams, it is clear that financing projects which rely on such sources to earn a return may be difficult. Volatility associated with merchant income is, in many jurisdictions, made worse by policy uncertainty. Policy choices drive the level of renewable investment, the extent of grid reinforcement and the extent of demand growth from newer flexible uses of power (such as electric vehicles), all of which influence market prices and the scope for time shifting and price arbitrage.

Set against the difficulty of financing BESSs is their importance to energy transition. As noted above, their ability to absorb excess intermittent renewable generation and provide a new source of synthetic inertia means they will be a fundamental part of any low carbon grid. In the UK, the “Clean Power 2030” plan for a low carbon grid foresees battery capacity between 23 and 27 GW. This may imply the need for further government intervention to support the investments.

As international electricity markets are gradually de-regulated, as is happening in many African jurisdictions, they may look overseas to historical precedents in deciding how to structure their markets and systems of government support. The UK (and to a lesser extent Europe) have historically been leaders in deregulation and electricity market innovation.

The possibility to provide support to low carbon sources of flexibility is explicitly foreseen in European legislation[6], and has been recognised in the UK government’s Review of Electricity Market Arrangements. However, there is less consensus on the design of an appropriate intervention.

In the UK, a cap and floor scheme is proposed for long duration energy storage (principally pumped hydro storage). The scheme would ensure that projects which the regulator recognises as beneficial receive a minimum level of gross margin. This floor is likely to be set at a level which ensures that reasonable levels of debt can be serviced. The quid pro quo is that the returns which plants can make will be capped. A similar regime is applied to interconnectors in the UK. The definition of “long duration” remains to be decided, but it is possible that some very long duration batteries may be eligible for this scheme.

In contrast, for shorter duration storage (more likely to be relevant for batteries), no specific scheme has yet been put forward. The government has indicated that it is considering modifications to the UK’s capacity auction arrangements. These see generators and storage operators offer to sell their availability to a central counterparty, and are designed to ensure that there is sufficient capacity on the grid to meet expected peaks in demand. At the moment the auctions are technologically neutral: fossil and non-fossil capacity competes in the same market (although there are already limits on the running hours of fossil fuelled plants).

But change may be in the air. While the final details are still being debated, the UK government might modify the auctions to ensure a minimum amount of low carbon flexibility is purchased. This would allow the price paid to low carbon flexible plants (such as batteries) to exceed that paid to other capacity (such as thermal generating plants). As with agreements concluded in today’s capacity auctions, the clearing price in such a modified auction would be indexed for 15 years and paid to investors by a central counterparty. The thinking is that this would again provide greater scope for debt financing.

Yet more variation is found in continental Europe. In Greece, the government has proposed a support regime for a pumped hydro plant (PHS Amfilochia). The effect of the arrangement would be that the plant’s investor would secure a regulated rate of return independent of merchant revenues. And in Italy, the Electricity Storage Capacity Procurement Mechanism (MACSE) also envisages provision of a largely regulated return to storage investors. In contrast to the Greek mechanism, the Italian regime would provide the potential for a small upside based on merchant returns.

As such, in looking to project finance BESSs in Europe, the scope for either long term bilateral contracts with blue-chip counterparties (e.g. to provide resilience of supply to datacentres) or policy support is likely to be an important factor in determining priority jurisdictions for investors, who will seek greater revenue and price certainty to underpin debt service and fixed operating costs and provide returns to equity. This is not to say that merchant projects are not possible. Many may still obtain financing, but only after careful diligence as to the likely evolution of merchant margins, and where stacking of revenues can provide some diversification and upside.

Key battery parameters and implications for financing BESS projects

In any discussion about structuring BESS projects and their financing, the particular properties and performance characteristics of batteries need to be taken into account.

Manufacturers of batteries define two key indicators which reflect their states and are useful in optimizing battery use and performance:

• Stage-of-Charge (SoC) is a measure which compares the current level of charge in the battery as a percentage of its level when fully charged, reflecting the quantity of electrical energy stored as a ratio of the maximum possible stored energy that the battery is capable of holding. As cell health declines the maximum possible charge that may be stored also declines. Another parameter that is sometimes referred to is Depth-of-Discharge (DoD), which is the inverse of SoC, so if the SoC is 80%, the DoD is 20%.
• State-of-Health (SoH) compares the maximum capacity of a fresh battery and a battery that has “aged” through use, owing to electrochemical deterioration. SoH is defined as the ratio of the maximum quantity of energy the battery is able to store at any time to its rated capacity, expressed as a percentage. As SoH degrades, the useable capacity of the battery diminishes because it will discharge sooner at a given rate of discharge (i.e. at a given output current).

Degradation in state of health (SoH)

The SoH of a lithium-ion battery declines with increasing number of battery charge and discharge cycles in a reasonably predictable manner, provided the battery is not excessively stressed. A typical rule of thumb is to assume a 10 year useable lifespan for daily charge/discharge cycles, i.e. around 4,000 cycles. However, at the upper end of the range, a well-known manufacturer’s sales literature indicates that its 68Ah cell reaches 80% SoH after 6,000 cycles,[7] representing a little over 16 years of daily cycles.

Cell lifespan may be affected by a number of factors including temperature, depth of discharge and charging current (C-rate); and achieving the upper end of the lifespan range may involve conservative assumptions about DoD and maintaining an optimum temperature within a tight range. SoH degradation curves may be non-linear and exhibit accelerated degradation with increasing number of cycles beyond a threshold point.

Useable lifetime and implications of degradation for system design

Usable battery lifetime (the point at which SoH has declined to a level which compromises the useability of a BESS) depends on the application. SoH degradation and the inherent decrease in capacity over time need to be taken into account in scoping and defining the services that a BESS project company commits to provide to an offtaker, as well as the duration of those services and the charges for those services.

If the project company’s contractual commitment is of sufficient duration, it may be necessary for it to incur capital expenditure to renew or add cells to restore the BESS’s performance; and this would be to be taken into account in calibrating service charges to be paid by the offtaker to the project company. The cost of renewing cells may however be difficult to predict; whilst cell costs have historically fallen over time, potential shortages in lithium and other essential raw materials and constraints on manufacturing capacity or increased demand might cause an unanticipated spike in prices.

In the case of a BESS that is routinely charged and discharged in daily cycles, the system lifetime and its economic life may be reasonably predictable. One example might be a BESS combined with a solar PV power plant that is charged during the daytime and discharged at night to provide power to (for example) a datacentre. Another example might be a BESS at an EV charging station which is charged during periods of low demand (at night) and discharged at times of peak EV charging demand – using such a system could relieve supply line congestion by spreading power supply demands over time.

However, in a more complex use scenario such as providing dynamic grid frequency stabilization services, the frequency of charge/discharge cycles may be more unpredictable as it may depend on more random factors such as wind speed variation/gusting. Consequently, the economic life of the BESS modules may vary widely and depend on the usage pattern.

In economic terms, an unpredictable usage pattern which may result in varying O&M costs (including capital expenditure being incurred at uncertain times to replace degraded cells) suggests a possible need to vary a portion of the charges for provision of the services according to the usage pattern (rather than merely levying a flat availability charge): this could be seen as analogous to the energy charge for a thermal power project which typically involves pass-through of variable operating costs that correlate with usage patterns.

Alternatively, the obligation to provide services could be inherently limited so that cell charge-discharge cycling constraints are respected and maintained within agreed limits.

Such factors are likely to be a key focus for potential lenders to a project who will be concerned if the project company is exposed to excessive risk in relation to the period over which the BESS is able to generate revenue and/or uncertainty over O&M costs. One option that might be considered by lenders is to sculpt the repayment schedule for their debt to take into account the rate of reduction in the system SoH to the extent that project revenues depend on the SoH and decline in tandem.

Battery warranties

Warranties are available from suppliers of batteries which guarantee their useable energy capacity (i.e. the SoH) for a defined period, typically up to ten years, based on defined usage parameters. Not surprisingly, the guaranteed capacity is related to the predicted SoH degradation curve, but it may be possible to modify the guarantee terms for a price so that they are more favourable.

One option that may be explored is an extended supplier’s warranty which artificially extends the SoH and slows the degradation curve for a fee – this is in effect a hybrid warranty/maintenance service as the supplier will inevitably have to replace degraded cells to achieve the extended BESS lifespan.

Mitigating degradation

The rate of degradation may be reduced by limiting maximum charge and depth of discharge (DoD) within a defined SoC window, which may be dynamically altered as the battery ages. Battery management systems may be programmed to manage SoC to increase lifespan at the expense of reducing useability. This is a common approach in EV battery management systems, preserving battery lifespan at the expense of maximum range.

Management of DoD and maintaining it within certain limits may also be required to preserve a valid manufacturer’s warranty or to achieve more favourable warranty terms. This has implications both for the technical design of BESS systems (and in particular battery management systems and software) and for scoping and defining the services that a BESS project company commits to provide to an offtaker and the technical limits of those services.

Environmental conditions

Cell performance and lifespan depend to a large extent on maintaining suitable environmental conditions. If the operating temperature is maintained within a relatively tight range, the cell lifespan may be enhanced and accordingly supplier warranties may be subject to specific environmental conditions being met and maintained such as maximum and minimum temperatures and limitations on the period of any temperature excursions.

In designing their systems, BESS operators will therefore need to consider how best to mitigate the risk of damage being caused to batteries or warranties being invalidated by thermal events, such as building in heating and ventilation system redundancy, incorporating back-up systems and modularization/containerisation of BESS units, so that in the worst case, only one module is compromised by any unplanned thermal excursion.

Other factors

Battery health is affected by charge and discharge rates (C-rates) but such limits should be built into the design of the battery management system and associated systems. Operators of BESSs will however wish to ensure that battery management systems (and their firmware/software) are capable of being supported over the long-term and that if they can no longer be supported, that they are readily able to be upgraded or replaced.

Second life batteries

Electric vehicle batteries that have reached the end of their usable life in mobile applications may have a second life in static BESS applications. The cost of such used batteries should be significantly lower than new batteries and their reduced energy density compared with new batteries might be less of a concern for stationary applications. For example, Nottingham City Council has installed 600kW of second life storage at its EV fleet depot to store excess electricity from on-site solar PV arrays which is then used to charge their EV fleet at peak times. The systems also aims to participate in grid services by trading stored electricity and providing vehicle-to-grid energy supply via bi-directional EV chargers.

Lenders may however have a concern about the remaining economic life of used cells and how predictable it is and may naturally be cautious and reluctant to take a view on the ability to replace such cells should they fail or reach an unacceptably low SOH. That said, the EV market is growing and the supply of used batteries should expand rapidly; and as the use of second life cells increases and the available data on their performance grows, the risks associated with such arrangements may become better understood and more predictable.

Conclusion

Battery energy storage systems represent a keystone for the transition towards a more sustainable energy generation and utilisation. Despite the value and advantages that they offer to enhance grid reliability and stability and to integrate with renewable power sources, significant challenges remain in securing financing for BESS projects. Addressing those challenges will require supportive regulatory frameworks, innovative government price and demand support arrangements, a flexible and innovative approach to project structuring, appropriate sharing of risk between operators and suppliers and technical solutions which mitigate commercial and technical risks. Overcoming these hurdles will allow the full potential of battery storage systems to be unlocked, paving the way for a more resilient and sustainable energy future.

[1] The U.S. Energy Information Administration records an average monthly round trip efficiency of 82% being achieved in 2019. The U.S. National Renewable Energy Laboratory 2021 Annual Technology Baseline figure is 86%.

[2] Data published by GivEnergy for its BESS products.

[3] For example, the UK’s National Energy System Operator has a licence obligation to maintain system frequency within a range of 50Hz +/- 1%, i.e. between 49.5 and 50.5 Hz.

[4] Wind turbines are built to be lightweight and have relatively low inertia. Variable speed wind turbines which utilise doubly fed induction generators (DFIGs) pose particular challenges: during a grid fault condition the power conversion system may be unable to handle the currents in both rotor and stator, leading to the wind turbine being disconnected from the grid.

[5] Even with high levels of synthetic inertia, a grid will still need “real” physical inertia. The UK National Grid ESO has contracted for several sources in the UK (e.g. at Deeside in England and Rassau in Wales) to provide “synchronous condensers” whereby motor-generators use grid power to spin up and maintain large masses in rotation to act as flywheels.

[6] For example, Article 19g of EU/2019/943 as amended, on “non-fossil flexibility support schemes”

[7] Lab test with 100% DoD, 1C/1C charge/discharge rate and at a temperature of 25°C.