Medium Duration Energy Storage – Kingpin of Net Zero Energy

We are at an interesting point in history where the entire shape of the future energy system is quite uncertain. On June 27^th of 2019, MP Chris Skidmore signed into law the UK’s commitment to net zero emission of greenhouse gas by 2050. In July 2019, a power auction in Portugal produced the lowest ever undertaking to produce electricity from photovoltaic (PV) panels at €14.76 /MWh. On September 20^th the third round of CfD (Contracts for Difference) auctions in the UK returned undertakings by a consortium comprising Equinor and SSE to deliver power to shore from a huge new offshore wind farm at the remarkable cost of £39.65/MWh (index linked to 2012) – less expensive than the average price paid presently to generators for electricity. Only four days later, EDF confirmed publicly that the Hinkley Point C nuclear power plant would cost some 15% more than more than had previously been thought. All of this happened against the backdrop of exponentially-rising societal awareness of the dangers of climate changed which propelled the “Extinction-Rebellion” demonstrations and the meteoric ascent of Greta Thunberg from unknown schoolgirl to Nobel Prize candidate. Nature itself was baring its teeth already before and during 2019 and the extensive Australian wildfires at the turn of this year together with double-flooding in the UK in February 2020 leave few rational beings still doubting that climate change is real and that energy usage has accelerated it.

If anything is certain, it must be that strategies developed prior to the middle of 2019 for how to achieve an affordable and environmentally-acceptable energy system for the UK are not now applicable. If illustration is required, the August 2018 report on “Analysis of Alternative Heat Decarbonisation Pathways” commissioned by the Committee on Climate Change to accompany its “2018 Progress Report to Parliament” uses figures of £40/MWh and £50/MWh as central assumptions for the levelised cost of energy (LCoE) in 2050 from PV and wind turbines respectively. Future energy strategies that were (at least arguably) completely defensible prior to 2019 pivoted around the tenets that (a) it is “essential” that a largescale infrastructure be established for carrying millions of tons of CO2 to sequestration sites, (b) significant “baseload generation” should be established using nuclear power and (c) that hydrogen will be produced in vast quantities and fed into our existing natural gas network to supplement or even replace natural gas as a heating fuel. The amazing cost-reduction progress made thus far by renewables has already opened up yawning sinkholes beneath these pillars and the engineering communities developing PV and wind generation are intent on going much further. 

The most established designs of wind-powered, solar-powered and nuclear-powered generation plant share one critically important feature contrasting them with the fossil-fuelled generation alternatives: These zero-carbon generators incur almost exactly the same total cost irrespective of what fraction of their total potential output is used. For example, a wind-turbine erected on the basis that it would receive £60 for each MWh produced will effectively cost £80/MWh if 25% of its potential energy output is turned-down and £120/MWh if 50% is turned-down. LCoE was a useful indicator when conventional forms of renewables and nuclear generation were providing only a very small proportion of total electricity production. In a future where it is certain that fossil-fuelled generation will diminish strongly, LCoE is no longer useful. In that future, the costs associated with reconciling the supply of electricity with the demand for electricity will be significant compared to the cost of primary generation. Collectively, these added system costs can be termed “flexibility costs”.  

There is already sufficient clear air between the LCoE of renewables and the next least expensive new generation options to prompt the obvious question: “Can flexibility costs be reduced to the level where the most attractive option for a future net-zero energy system draws all primary energy from wind, solar and nuclear sources?”  If energy storage technologies had sufficiently low cost and sufficiently high performance, the answer to this question would obviously be “Yes”.

Arrangements were begun in December 2019 by Prof. S. Garvey of the University of Nottingham for a one-day meeting that was to be held on March 23^rd of 2020 at the IMechE headquarters at One Birdcage Walk. Speakers were enrolled and delegates were invited to address the topic of “Medium Duration Energy Storage in the Net Zero UK”. COVID-19 precluded the conduct of the event as a physical meeting but the event proceeded very satisfactorily as an E-meeting. Energy storage was defined very broadly for the purposes of this meeting to include storing energy prior to the generation of electricity and the storage of energy services such as heat, coolth and compressed air.

 According to Garvey, “One strong theme introduced at the start of the workshop and reinforced by every single contributor was this: In the context of future net-zero energy systems, energy storage will be required over a vast range of discharge times spanning <0.1 seconds up to >12 months – a dynamic range of more than 100 million (or >28 octaves to give an acoustic / musical measure). No one single set of technologies is suited to deal with this complete range of discharge times and there are strong arguments to support distinguishing four different main ranges: (I) very-short duration storage (<5 minutes) which is handled best by flywheels, supercapacitors and possibly large inductors, (II) short-duration storage (5 minutes – 4 hours) which is dominated by electrochemical batteries and demand-side response actions, (III) medium-duration storage (4 hours – 200 hours) where thermo-mechanical solutions comprise the main options and (IV) long-duration storage (>200 hours) which is mainly achieved by storing fuels such as bio-mass, bio-gas, ammonia, other synthesised fuels and hydrogen.”

The figure below was presented to illustrate.

Garvey asserts that “Present energy policy focuses attention almost exclusively to categories (II) and (IV) and yet there was consensus that the medium-duration energy storage (category (III)) would do the heavy-lifting in the sense that the bulk (>90%) of all energy emerging from storage would be from stores in this category. Correspondingly, it is likely that the greatest capital spend on energy storage infrastructure in the future should and will be on facilities in this set. Policy needs to change on this important point.”

There were two presentations addressing the drivers for energy storage in Northern Europe coming specifically from the huge potential of offshore wind. Both emphasised that the offshore wind resource was more than sufficient to meet all energy needs and both recognised that exploiting even a part of this resource would indeed raise major requirements for medium-duration storage. Henrik Stiesdal, a globally-recognised thought-leader from the wind turbine industry and former CTO of Siemens Gamesa Renewable Energy, elucidated this point for the particular case of a future Danish energy system but his reasoning extends naturally to all of Northern Europe.

It was highlighted several times that by far the largest energy storage capacity would need to be provided in the long duration stores. There is a common misunderstanding that the seasonality of energy demand is the main driver for long duration storage and this is quite incorrect. Wind power, like demand, varies strongly with season and both are strongest in winter. By contrast, solar power is seasonal and obviously peaks strongly in summer. If we could rely on the monthly averages of wind and solar power being similar from year to year, there would scarcely be any need for long duration storage because a generation mix of ~80% wind and ~20% solar would roughly match present seasonality of demand. The key driver for this category of storage lies in the strong variation between years. The long duration energy storage facilities need have only relatively small power ratings. The reasons why long duration storage cannot subsume the duty best served by the medium duration stores is two-fold: (i) the effective turnaround efficiency of these stores is relatively low and (ii) the cost of the power-conversion equipment (£/MW(rated)) is relatively high.

Another point echoed multiple times was that a small amount of overcapacity in total renewable generation is very cost-effective overall because of the corresponding reduction that occurs in energy storage requirements. Future systems that generate 10 – 15% more electricity from renewables in a year than is expected to be consumed during the average year appear to be optimal based on present cost assessments. If primary generation costs continue to reduce, then higher proportions of over generation will be appropriate.

Three separate presentations considered the storage of heat (especially in dedicated pits filled with water), the storage of coolth (exploiting phase-change materials among other options) and exploiting the liquefaction of air to enable, in effect, large amounts of compressed air to be stored at relatively low cost.

To the majority of people, “Energy Storage” has become synonymous with electrochemical batteries. One of the messages emerging from the meeting was that the energy storage market naturally begins with short-duration storage and at present, this is the only market on which commercial returns are possible. Batteries suit this market very well so the common understanding is entirely appropriate at the moment but several of the presenters at the meeting were at pains to emphasise that as the penetration of low-carbon generation increases, the discharge durations required also get longer. It was also noted that whilst batteries are naturally well-suited to being developed spontaneously by private investment, this will not work for the medium duration and long duration stores where (a) the natural lifetimes of the facilities created far exceeds the investment horizons of private investors and where (b) good economics and good performance depends intrinsically on realising individual installations at very large scales.

Professor Sir Chris Llewellyn Smith FRS of Oxford University was the natural choice to provide the keynote address launching the event because he is presently leading a detailed review on behalf of the Royal Society into energy storage over long periods. That review will release its final report later this year. Prof. Llewellyn-Smith used data over a 37 year period to reveal huge variability between years in the total annual energy production that would be achieved by any given fleet of wind turbines in the UK or any fixed collection of PV panels. This remarkable observation justifies both a degree of net over-capacity in renewable energy for the average year and points to a future requirement for enormous quantities of fuel storage – equivalent to several full weeks worth of present average electricity consumption.

The chairman for the day was Professor Paul Ekins from Bartlett School of Environment, Energy and Resources at UCL. Prof. Ekins noted “It is clear that medium-duration energy storage will be needed to play a critical in making the most of renewable power, and there are numerous promising options. They need to be demonstrated and deployed at scale to see which of them can make breakthroughs in terms of performance and cost.”

Gareth Brett, CTO of Highview Power Ltd., gave the presentation about liquid air energy storage. He commented that “Storage technologies with greater than 4 hours duration are a key enabler for the net zero energy system and I’m delighted to see a community coalescing around this view. In our development work we’ve also found that scale and locatability are important factors to ensure that the capabilities and benefits of storage can be accessed by the transmission network and value transferred to the owners of storage (thereby justifying the investment). These are features designed into Highview’s liquid air energy storage systems giving access to pumped hydro performance and capability without any geographical constraints”.

The event was very well attended on the day by industry, policy-makers and academe alike. It ended with a panel session including substantive contributions from the IMechE’s own Jennifer Baxter, who emphasised the need to consider the complete energy system including heating and transport, and National Grid’s Head of Networks Julian Leslie who pointed to the need for a shift from “dispatchable generation” towards “dispatchable demand”.  It was also fully recorded so that those unable to join at the time would have access to the materials afterwards. The complete transactions, comprising videos of the presentations themselves as well as the slide-decks, are available without charge by visiting the download page on the Energy Research Accelerator website.