SAVECAES: Sustainable, Affordable and Viable Compressed Air Energy Storage – Written evidence (LES0024)


Principal Investigator: Dr E Barbour, Loughborough University

SAVECAES: EPSRC Grant EP/W027569/1


List of acronyms used:

AC                            Air Conditioning

CAES                            Compressed Air Energy Storage

CCS                            Carbon Capture and Storage

DAC                            Direct Air Capture

DNO                            Distribution Network Operator

ESO                            Electricity System Operator

LAES                            Liquid Air Energy Storage

PCM                            Phase Change Material

PEM                            Proton Exchange Membrane

PTES                            Pumped Thermal Energy Storage

SMR                            Steam Methane Reforming

TRL                             Technology Readiness Level



How much medium and long duration energy storage will be required to reach the government’s goal of a fully decarbonised power grid by 2035 and net zero by 2050, and when will it need to be ready?


A reasonable estimate is 20-50 TWh of long duration storage (primarily to meet seasonal heat demands) and 200-500 GWh of medium duration storage (to manage large swings in wind generation). However, these numbers depend on the mix of technologies adopted - principally the balance of nuclear, wind and solar generators as our main low-carbon energy generation options. Of the main Net Zero pathways currently in existence, long duration storage is typically in the range 20-50 TWh (hydrogen or another seasonal fuel) and medium duration storage comprises 100’s of GWh (compressed air, flow batteries, pumped hydro)[1]. Only scenarios which rely on huge successes in CCS (Carbon Capture and Storage) and Direct Air Capture (DAC) technologies, which remain stubbornly unproven despite huge investment, do not require storage at this level.


An illustrative, back-of-the-envelope example is to consider 1 week with low wind in 2050: Assuming one week at 10% capacity factor for a system with 100 GW of wind, sized for approximately 40% capacity factor, gives a shortage of approximately 5 TWh. In fact this level of wind is extremely conservative without CCS or DAC (both of which remain unproved) and is likely to at least double with a majority of transport and heat electrification. Energy conversion losses in the storage also increase this 5 TWh figure.


We need to be funding pilot systems now in order to have the first generation of major pilot plants under construction in the next 2-5 years. Later than this, we will not have the engineering knowledge to build out the required capacity in time, and we will also lose the internationally-leading positions we currently enjoy. Given the long lead times required for nuclear deployment, the demand for energy storage will be shifted closer to the current time than would be the case if we had more nuclear power currently under construction. Furthermore, at present nuclear costs cannot compete with solar or wind, and both can be deployed much faster, at the expense of increased intermittency. By 2035 it is likely we will need around 10-20% of storage in place, with the required capacity rising rapidly as the pace of electrification in the heat and transportation sectors increases from 2035 to 2050. Electrification of heat will have a particularly strong contribution to the amount of medium and long duration storage required.


How sensitive is the amount of storage needed to assumptions about the future balance of supply and demand on the grid


The absolute amount of storage needed is certainly sensitive to the balance of supply and demand. Nuclear generation in 2050 tends to reduce the amount of storage required by providing baseload, as does finding the optimal balance of wind and solar (it is generally felt that for the UK, the optimum mix is ~85% wind and ~15% solar, by total annual generation). However, the only scenarios in which we do not need very large amounts of storage are those which assume massive improvements in CCS, including DAC.


The inclusion of demand response also reduces energy storage requirements, however this only displaces short term storage, since demand response is unlikely to be able to shift energy use by more than a few hours. Interconnectors also provide flexibility, but this again is typically limited to displacing short term storage. To displace medium and long duration storage, demands for heat and transport must be substantially reduced over sustained periods. The only suggested ways to do this at present without storage include: (i) Significant reductions in indoor air temperatures tolerated in winter in UK buildings. (ii) Very significant improvements in building fabric performance. (iii) Big modal shifts in transportation to mass public transport or walking/cycling.


There is also some uncertainty regarding the impact of climate change. This could lead to warmer winters which would reduce heating requirements, however the likelihood of sustained cold over periods of a couple of days will persist. Thus winter heating loads are the largest contributor to storage needs. Air Conditioning loads seem unlikely to have a huge influence on storage needs in the UK as: (i) They align well with peak solar production and hence can be met with short duration storage. (ii) There is little evidence yet that AC will be widely adopted in residences.


The uncertainty described in the previous paragraphs highlights a broader, and crucial point, about how much energy storage we need. We can only take seriously the estimates of required energy storage capacity that use models which are open-source and transparent. It is relatively easy to dream up a model for the UK in 2050 with little storage, which meets net zero goals and looks attractive from a cost perspective. However, when scrutinised these models always have questionable assumptions and optimistic costs for technologies which are not even close to demonstrable. Therefore, any modelling work on which we make our decisions must be transparent and open to scrutiny.


Which technologies can scale up to play a major role in storage?


The key trait for medium and long duration energy storage technologies is an ability to decouple the power component of the storage from the energy component. In this regard, technologies like Compressed Air Energy Storage (CAES), Pumped Thermal Energy Storage (PTES), Liquid Air Energy Storage (LAES) and flow batteries are all promising for medium duration energy storage. For long-duration storage, the most suitable technologies are hydrogen, synthetic methane, ammonia or other synthetic fuels that can be stored seasonally. Thermal energy storage represents a suite of technologies with various options to cover both medium duration (i.e. sensible heat stores, phase change material stores) and long-duration storage (i.e. underground water heat stores, thermochemical stores). All of these technologies have the potential for major roles, however they are currently at different stages of development.


Of the options for medium duration energy storage, CAES is the most mature, although it remains unproven at scale for versions of the technology that do not require supplementary fossil fuel use. Next-generation CAES can build on the learning from large-scale CAES systems (using fossil fuel) already in existence, noting that the engineering related to the gas storage does not change. In general, the cheapest variant of this technology uses underground salt caverns to store the high-pressure air and the UK has an excellent salt cavern resource. Typical volumes of a CAES salt cavern are around 100,000 cubic metres, which yields a plant size of around 400-1000 MWh. For reference, this is ~10X smaller than the capacity of the Dinorwig pumped hydro plant, however unlike pumped hydro, it is estimated that the UK could house hundreds of these caverns in the Cheshire basin alone, leading to in excess of 2 TWh of medium duration energy storage[2].


For long duration energy storage, hydrogen storage in salt caverns also certainly has potential and is a proven technology. However, generating hydrogen without Steam Methane Reforming (SMR) remains prohibitively expensive at the present time, and has limited potential for further cost reduction without the development of electrolysis methods without iridium catalysts. This problem is common with ammonia and synthetic methane since both of these options require hydrogen.


Given the role of heat demand in driving the need for medium and long duration storage, thermal storage development should be prioritised. Thermal storage also has the major attraction that it could be deployed in individual buildings as a modular solution and that, if deployed at the individual building level, it could significantly reduce the peak electrical power demands on the network. The UK has the research capability in place to develop modular thermal energy storage solutions for households, including sensible, PCM and thermochemical systems and this capability should be exploited.


What policy support is currently in place to support deployment of storage technologies? Is it sufficient to support deployment at scale?


The empirical evidence is clear - the policy support for the development and subsequent deployment of medium and long duration energy storage is not in place. Small schemes to support medium and long duration storage, such as the recent BEIS Longer Duration Energy Storage Demonstration (LODES) Competition are undoubtedly useful, but do not go far enough. The government must make its commitment to long duration storage clear. A mandate that major utilities procure a minimum level of longer duration energy storage would send this message clearly. This would force the development of long duration energy storage and move the incentive in storage development away from being able to respond over short timeframes, for which we already have suitable technologies. It would also provide a major incentive for knowledge surrounding medium and long duration energy storage systems to be brought together, since at present it is fragmented across large utilities, early-stage ventures and academia. Without direct mandates, market actors are incredibly unlikely to deliver storage needs, given the aversion to long-lifetime, adventurous, high-capital-cost projects.


How well developed is the UK industry across different storage technologies, such as hydrogen or redox flow batteries? How does the UK compare to global competitors in these industries?


Through its world-class universities, the UK is a global leader in CAES, PHES, LAES and thermal energy storage solutions. It also has a very strong and historical mechanical engineering pedigree, covering relevant and complementary areas such as combustion engines, aero-derivative engines and turbomachinery. UK companies such as Rolls Royce are well positioned to aid development in CAES and other major employers in similar industries in the UK (such as GE) mean that there is a highly skilled workforce available. In CAES, the UK competes with China to be a leader in next generation systems, with China having the edge due to significant expenditure in large-scale pilot projects over the past 10 years. However, the strength of CAES research in the UK, combined with the available industrial expertise in complementary areas, mean that any funding towards large-scale pilot projects would likely be fruitful. CAES also has the potential to build on existing infrastructure available in the UK, including disused caverns, mines and natural gas pipelines, as well as expertise in solution mining of salt caverns and a considerable salt cavern resource.


In lower TRL level medium duration energy storage technologies, such as PTES, the UK is also the global leader. Highview Power is the global leader in LAES, although the academic consensus is generally that the efficiency of LAES plants is lower than other long duration options. Hence, it is most likely to be usefully deployed for niche applications where waste heat sources are available in close proximity.


In green hydrogen development, ITM power is a global producer of PEM electrolysers, and as mentioned for CAES, the UK potentially has a large resource of underground salt caverns which could facilitate hydrogen storage.


Beyond the cost of deploying long-duration energy storage, what major barriers exist to its successful scale up (e.g. the availability of a skilled workforce, the ability to construct the necessary infrastructure on time, or safety concerns around new technologies)?


The motivation behind medium and long duration energy storage must be clearly articulated across academia, industry and government. This means that it should be underpinned by open-source and transparent models that stand up to scrutiny. Conflicting results delivered by opaque modelling strategies already confuse the debate and there is a major danger that unless consensus open-source work becomes available, we will not make the correct decisions regarding storage until it is too late.


The ability to construct infrastructure on time is a major concern. Long duration energy storage projects must be given priority for grid connections, i.e. assurances that when they are completed they will be connected to the grid with the highest priority. This, of course, makes sense since storage can always be operated in a manner to aid with grid constraints. At the current time of writing, renewable generation projects are receiving connection times in 2040. The process for grid connections needs to be reformed, with all projects that have not started construction removed from the queue.


Safety around new technologies is also of paramount importance, however, it should be recognised that the risks with many medium duration energy storage technologies - including CAES, PTES, LAES and thermal energy storage - are comparatively low compared to hydrogen and electrochemical options at large-scale. The supply chain for these technologies also does not typically include rare and exotic materials, which means smaller environmental impacts and also independence from external suppliers.


As with several new and novel energy technologies, the role of the non-expert stakeholders (including the general public, those living locally to proposed schemes) in schemes gaining a “social license to operate” should be carefully considered. The impact where public acceptance becomes problematic can include delays to planning/construction, cancellation of schemes and general stress to communities. Studies examining how best to inform communities and non-experts could be commissioned to enable a better understanding of how long- and medium-duration energy storage are understood and accepted by communities impacted by developments.


One of the biggest barriers to deployment and scale up is the lack of any mandates for long duration energy storage. A mandate for the big energy generators to develop pilot systems at 10-50 MW scale, with eligible systems able to produce >8-12 hours of output at full discharge power, would be hugely beneficial. It would also swiftly facilitate a transition from lab-scale experiments to real-world development.


What steps should the Government take now to ensure this storage can come online later in the current decade?


HM’s government must propose a clear and strict set of directives to guide research and commercial ventures to develop long duration storage. This could either be through public sector commissioning or a mandate regarding storage investment or procurement.


A dedicated, and independent, body should be formed to understand the TRL of the various medium and long duration energy storage technologies. This body should then have the authority to arrange for key demonstration funding rounds and be able to monitor performance. The body should also specify a clear set of metrics for comparing long duration energy storage technologies, so that early-stage developers are clear on what they are aiming to achieve. This would also avoid developers aiming at current revenue mechanisms which are; (i) unlikely to be the dominant modes for storage revenue in the future, and (ii) result in suboptimal designs for long duration storage. This body must be wholly independent from the current ESO and DNOs to avoid constrained thinking and should be able to direct funding to pilot plant proposals in the most promising technologies.


All decision making must be transparent and there should be a public release of modelling frameworks, including timeseries of supply and demand. This would be invaluable for storage developers seeking to understand how their technology can be implemented within a future net zero energy system.


11 September 2023


[1] Scenarios included - National Grid FES Leading the Way, System transformation, Consumer Transformation and Centre for Alternative Technology Zero Carbon Britain. Only the Energy Systems Catapult scenarios have significantly less medium and long duration storage, rather relying on CCS, including DAC. Summary of net zero scenarios.

[2] King, Marcus, et al. "Overview of current compressed air energy storage projects and analysis of the potential underground storage capacity in India and the UK." Renewable and Sustainable Energy Reviews 139 (2021): 110705.