The increasing importance of electricity storage in high variable renewable electricity systems

KEY POINTS:

As the penetration of variable renewables increases, so the benefits of storage increase

Electricity storage is one of the most flexible options to increase electricity system flexibility, because of its ability to produce and consume electricity, in contrast to demand-side response (DSR) or flexible generation, and to shift electricity supply in time, in contrast to interconnection. The globally installed electricity storage capacity is projected to almost double from 145 GW in 2010 to 266 GW by 2030 [1].

Electricity storage is already making significant contributions to electricity system reserve services.

In Great Britain, for example, the increased need to stabilize the electricity system short-term due to an increased VRE share and fewer centralized power stations providing system inertia has led to the establishment of a new service, Enhanced Frequency Response, with the initial requirement of 200 MW taken entirely by battery technologies [2]. Similarly, increased supply uncertainty due to VRE combined with a delicate power network, leading to blackouts in South Australia, triggered the installation of a 100 MW battery that outperforms traditional alternatives in terms of accuracy and speed to stabilize the power system in the short-term, as well as the number of different reserve and balancing services it can provide [3]. Information on revenues indicate that the project will recoup its investment cost within 2 years [4]. Reducing feed-in incentives for VRE and high network expansion costs has led to increased interest in use cases that allow the time-shift of VRE to benefit from time-of-use tariffs or changes in wholesale prices or defer investments in network infrastructure [5].

Multiple electricity sector applications can be served by electricity storage technologies.

There is a multitude of power system services required to balance electricity supply with demand due to the transformation of the power system as outlined above. These services differ in purpose and technical requirements such as size, response time, duration and number of events (see Figure below) [6–8]. For example electricity arbitrage, whereby electricity can be bought and stored when cheap and plentiful, and supplied when scarce and expensive, does not require a fast response, but rather large systems that can store electricity for many hours, possibly once a day. In contrast, frequency regulation requires a very fast response to provide electricity for only minutes, several times a day. Customer-side backup power may only be required a few times per year, but for as long as the outage lasts, which could be several hours.

Figure 1 – System Operator, Utility and Customer services that electricity storage technologies can provide at different locations of the power system [8].

And there is a variety of storage technologies, each suited to different applications.

Similarly, there is a multitude of energy storage technologies that differ in underlying physical concept (electric, electro-magnetic, electro-chemical, mechanical, chemical, thermal) and resulting differences in cost and performance parameters.

Some electricity storage technologies like capacitors and coils are relatively small and respond and charge / discharge ultra-fast. By contrast, mechanical, thermal and chemical storage technologies like pumped hydro, district heat storage or power-to-gas, respond slower, have large storage capacities and can provide electricity for many hours or even longer. Electrochemical storage technologies (i.e. batteries) tend to lie between these extremes, responding very fast and featuring moderate electricity storage capacities and discharge duration's.

The challenge is in matching technologies and applications by optimizing technology cost and performance parameters to application requirements, while accounting for the rapid cost reductions experienced by certain technologies.

Investment costs of storage technologies are coming down, in some cases very fast.

Many electricity storage technologies are experiencing significant investment cost reductions due to R&D investments, economies of scale, and manufacturing experience. In particular, lithium-ion (li-ion) batteries benefit from the production scale-up experienced since the 1990s for consumer electronics applications, and since 2010 for electric vehicle battery packs, with total cost reducing from 3,000 US$/kWh to just above 100 US$/kWh for a li-ion battery cell. These cost reductions make the technology a promising candidate for stationary storage applications. Since the battery cells constitute only a small part of a stationary battery system, stationary lithium-ion systems including balance-of-plant components, the power control system, as well as engineering and procurement, cost around 1,000 US$/kWh but are expected to reduce to below 500 US$/kWh by 2030 [9]. Other novel battery storage technologies have a similar potential for cost reductions but need a sizable market for realization. Established technologies like pumped hydro storage plants or lead-acid batteries are not expected to benefit from future investment cost reductions but remain competitive due to already low costs. Performance requirements or geographic constraints may hinder their deployment and drive interest in cost reductions of alternatives.

It is important to consider not just investment costs, but full levelized costs.

The complexity of electricity storage cost and performance parameters and different application requirements make it challenging to identify the most economic storage solution. The Levelized cost of storage (LCOS) accounts for all technical and economic parameters affecting the lifetime cost of discharging stored electricity in a specific application, and therefore represents an appropriate tool for comparing the effective cost of storage technologies [10–12].

The choice of electricity storage technologies must be appropriate for the required size of the system, the required electricity discharge duration and the discharge events per year. These application requirements affect LCOS and thereby determine technology suitability (Figure 2).

Battery and thermal technologies are most cost-effective for residential / commercial applications, with lithium-ion and lead-acid most suitable for short discharge durations and thermal, vanadium redox-flow and sodium-sulfur for long discharge durations with many cycles, outcompeting lithium- ion that currently cannot sustain comparable cycles throughout its lifetime. Lead-acid is most cost- effective for long discharge durations with relatively few cycles. By 2030 and 2040, current cost projections suggest that the economics of vanadium redox-flow and particularly lithium-ion will improve and make them most suitable for all residential and commercial applications.

Electrical technologies (magnetic coils and super-capacitors), flywheels and lithium-ion batteries are most cost-competitive for industry / utility applications with many cycles and short discharge durations. The main criteria here are low investment costs per unit of power output, and a high cycle life. Projecting forwards, lithium-ion is again likely to become the most dominant technology.

Mechanical, chemical and thermal storage technologies are more suitable for long discharge duration applications due to their relatively low investment costs per unit of electricity stored. Again, the substantial investment cost reduction projected for lithium-ion batteries, and potentially also vanadium redox-flow batteries, could challenge this dominance. Vanadium redox-flow has a good cycle life and its electricity storage capacity can be increased independently from its power output (which is not the case for lithium-ion), allowing economies of scale for large electricity storage capacity systems. However, without a sizable current market, future investment cost reductions are less certain than for lithium-ion.

Figure 2: Suitability of energy storage technologies to energy system applications based on analysis of lowest levelized cost of storage for 2030 [10-12]. Assessment of thermal storage based on industry trend, not LCOS calculation.

Power Quality: This service is characterized by discharge duration's below one hour per event and up to a couple hundred events per year. Lithium-ion, sodium-sulfur and lead-acid batteries are most suitable due to balanced cost for specific energy storage and power provision capacity and sufficient cycle life. Super-capacitors, magnetic coils and flywheels have too high energy cost. Vanadium redox- flow and power-to-gas have too high power capacity costs. Future investment cost projections increase competitiveness of lithium-ion and vanadium redox-flow. Pumped hydro, compressed air and thermal storage systems are too large customers typically requiring this service.

Demand charge reduction: Albeit a different purpose, this service has very similar performance requirements to power quality. Hence, technology suitability follows the same rationale.

Time-of-use bill management: This service is characterized by a couple of hours discharge per event and a multiple hundred events per year. Lithium-ion, vanadium redox-flow and sodium-sulfur are most suitable due to relatively low energy storage capacity costs and sufficient cycle life. Lead- acid has too short cycle life to be competitive. Projected cost reductions mean that lithium-ion and vanadium redox-flow likely out-compete sodium-sulfur. Pumped hydro and compressed air too large for customers requiring this service. Supercapacitors, magnetic coils and flywheels are ill-suited for this service due to cost per electricity storage capacity.

Solar PV self-consumption: Albeit a different purpose, this service has very similar performance requirements to time-of-use bill management. Hence, technology suitability follows the same rationale.

Backup power: This service is characterized by multiple hours discharge per event, but only a few events per year. Lead-acid batteries are most competitive due to low specific energy storage capacity cost and sufficient cycle life for few annual events. Other battery technologies are also competitive and future investment cost projections leave lithium-ion and vanadium redox-flow as most competitive for this application. Pumped hydro, compressed air and thermal storage systems are too large customers usually requiring this service. Supercapacitors, magnetic coils and flywheels are ill-suited due to high specific energy capacity costs.

Frequency regulation: This service is characterized by short discharge duration below one hour per event and hundreds to thousands of events per year. Supercapacitors, magnetic coils and flywheels are competitive due to their low specific power provision capacity cost, very high cycle life and round-trip efficiency. Lithium-ion batteries are also competitive in situations where long discharge duration systems are required (>0.5 hours) due to their much lower specific energy storage costs. Projected investment cost reductions increase the competitiveness of lithium-ion and flywheels. Pumped hydro, compressed air and thermal storage systems have too slow response time for this service and alternative battery technologies either have too low cycle life or too high specific power provision capacity costs.

Voltage support: Albeit a different purpose, this service has very similar performance requirements to frequency regulation. Hence, technology suitability follows the same rationale.

Network congestion relief: This service is characterized by a couple of hours discharge per event and hundreds of events per year. Pumped hydro and compressed air storage are most competitive due to low specific energy capacity and moderate specific power provision capacity costs and good cycle life. Pumped hydro ends up more competitive than compressed air due the higher round- trip efficiency, which matters at a high number of events per year. Projected cost reductions make lithium-ion and vanadium redox-flow more competitive due to lower specific power capacity costs, outweighing shorter cycle life.

Black start: This service is characterized by only up to a few hours discharge per event and few events per year. Pumped hydro, compressed air and lead-acid battery storage are most competitive due to low specific energy storage and moderate specific power provision capacity costs. Projected cost reductions make lithium-ion more competitive due to lower specific power provision capacity costs. The lower cycle life of battery technologies compared to pumped hydro and compressed air is not significant in this application due to the few events per year.

Spin / Non-spin reserve: This service is characterized by up to an hour discharge per event and up to a thousand events per year. Pumped hydro storage is most competitive, followed by flywheels due to the potentially relatively high cycle requirement. Compressed air storage is negatively affected by its low round-trip efficiency. Investment cost projections indicate that lithium-ion and vanadium redox-flow could become contenders for this service.

Energy Arbitrage: This service is characterized by multiple hours of discharge per event and a couple of hundred events per year. Pumped hydro storage is most competitive due to low specific energy storage capacity costs combined with high round-trip efficiency and cycle life. Compressed air storage is negatively affected by its low round-trip efficiency. Investment cost projections indicate that lithium-ion and vanadium redox-flow could become contenders for this service.

Network investment deferral: This service is similar to energy arbitrage, albeit fewer cycles per year. Hence, compressed air could be as competitive as pumped hydro energy storage. Investment cost projections indicate that lithium-ion and vanadium redox-flow could become contenders for this service.

Seasonal storage: This service is characterized by very long discharge per event and few events per year. It is only relevant for technologies that can scale energy storage capacity fully independently from power capacity. Initially, pumped hydro, compressed air, power-to-gas and thermal storage technologies are most competitive due to their very low specific energy storage capacity costs. Cost projections indicate improving competitiveness of power-to-gas.

References

[1]   Tracking Clean Energy Progress 2017, (2017), International Energy Agency.

[2]  National Grid, (2016), Enhanced Frequency Response - Market Information Report.

[3]  AEMO, (2018), Initial operation of the Hornsdale Power Reserve Battery Energy Storage System.

[4]  Parkinson G. Revealed: True cost of Tesla big battery, and its government contract | Renew Economy. Renew Econ n.d.

[5]  DNV GL, (2016), Energy Storage Use Cases.

[6]  Huff G, Currier AB, Kaun BC, Rastler DM, Chen SB, Bradshaw DT, et al, (2013), DOE/EPRI 2013 electricity storage handbook in collaboration with NRECA. Sandia National Laboratories; doi:SAND2013-5131.

[7]   IEA, (2014), Technology Roadmap - Energy Storage. Paris: International Energy Agency;

[8]  Fitzgerald G, Mandel J, Morris J, Touati H, (2015), The Economics of Battery Energy Storage: how multi-use, customer-sited batteries deliver the most services and value to customers and the grid.

[9]  Schmidt O, Hawkes A, Gambhir A, Staffell I. The future cost of electrical energy storage based on experience curves. Nat Energy 2017;2:17110. doi:10.1038/nenergy.2017.110.

[10] Lazard, (2017), Lazard’s Levelized Cost of Storage Analysis - Version 3.0 [11] Lazard, (2016), Lazard’s levelized cost of storage analysis - version 2.0

[12] Schmidt O, Melchior S, Hawkes A, Staffell I, (2018), Projecting the future levelized cost of electricity storage technologies.