Adiabatic compressed air energy storage technology

Abstract

Graphical AbstractView Large Image Figure ViewerDownload Hi-res image Download (PPT)For decades, technical literature has appraised adiabatic compressed air energy storage (ACAES) as a potential long-duration energy storage solution. However, it has not reached the expected performance indicators and widespread implementation. Here, we reflect on the design requirements and specific challenges for each ACAES component. We use evidence from recent numerical, theoretical, and experimental studies to define the technology-readiness level (TRL). Lastly, we discuss promising new directions for future technology development. For decades, technical literature has appraised adiabatic compressed air energy storage (ACAES) as a potential long-duration energy storage solution. However, it has not reached the expected performance indicators and widespread implementation. Here, we reflect on the design requirements and specific challenges for each ACAES component. We use evidence from recent numerical, theoretical, and experimental studies to define the technology-readiness level (TRL). Lastly, we discuss promising new directions for future technology development. Adiabatic compressed air energy storage (ACAES) is frequently suggested as a promising alternative for bulk electricity storage, alongside more established technologies such as pumped hydroelectric storage and, more recently, high-capacity batteries, but as yet no viable ACAES plant exists. At first sight, this appears surprising, given that technical literature consistently refers to its potential as a promising energy storage solution and the fact that two diabatic compressed air energy storage (DCAES) plants exist at utility scale (Huntorf, Germany and Macintosh Alabama, USA), with over 80 years of combined operation. In this article, we discuss aspects of the main components that constitute a compressed air energy storage (CAES) system, the fundamental differences between how they operate in diabatic and adiabatic contexts, and the design challenges that need to be overcome for ACAES to become a viable energy storage option in the future. These challenges are grounded in thermodynamics and are consistent with evidence from pilot plants, where performance information has been made available. Finally, we suggest that adopting a whole systems design philosophy would maximize the chances of successful ACAES demonstration and discuss worthwhile areas of future research for both simulation and experimental studies. Any CAES system is charged by using electricity to drive air compressors, resulting in compressed air and heat. In DCAES, the heat is extracted by using heat exchangers (HEX) and dissipated (being of low grade and therefore of low value), whereas the pressurized air is stored in a dedicated pressure vessel, herein referred to as the high-pressure (HP) store. Crucially, there is no heat transfer between the charging and discharging processes. To discharge the system, air is released from the HP store and heated by a combustion chamber by using natural gas and is finally expanded through turbines generating electricity. This process is illustrated in the lower section of Figure 1. The process for ACAES is conceptually similar to DCAES (Figure 1, upper section) but has different heat management. In ACAES, the heat from the compression is stored and is used to reheat the air prior to expansion, removing the need for additional fuel but necessitating a thermal energy store (TES). However, whereas DCAES and ACAES might be similar to some degree, the need to store the compression heat dramatically changes the design of nearly all of the common constituent components. Equation 1 gives the isentropic work, ΔW, which is required to compress an air mass Δm from some inlet pressure and temperature, pi and Ti, respectively, to some outlet pressure po in a single adiabatic compression stage:ΔWΔm={ho−hicpTi[(popi)γ−1γ−1]perfectgasapproximation.(Equation 1) Here, ho−hi denotes the change in enthalpy from the inlet-to-outlet air state, cp is the specific heat capacity at constant pressure, dh=cpdT with the perfect gas approximation (simplifying the ideal gas model by assuming constant specific heats), and γ is the ratio of specific heats (γ≈1.4 for dry air at standard conditions). This process is isentropic and reversible, and accordingly, Equation 1 also gives the work available from air expansion (available work has the opposite sign). Studying Equation 1, several aspects of CAES become apparent. First, for both DCAES and ACAES, unless the pressure in the HP store is controlled to be constant, or there is significant overcompression, the final compression outlet pressure will be variable. This increasing outlet pressure during charging also leads to variable outlet temperatures, and the effect is mirrored during discharging. Second, the compression and expansion processes are strictly coupled in ACAES because without a supplemental heat input, the full compression work can only be recovered when the expansion inlet temperature is the same as the compression outlet temperature (i.e., Tie=Toc=(popi)γ−1γ). Therefore, for maximum efficiency, the expansion process must reverse the compression path and the TES should preserve the compression temperatures. This is not true in the case of DCAES, as the external heat source can independently supply heat at a specified temperature, decoupling the charging and discharging processes. Third, the work required to compress a given mass of air is minimized by staging the compression and cooling the air back to ambient temperature between stages (minimum work is attained with equal pressure ratio in all stages). In the limit of perfect intercooling and N stages, Equation 1 becomes ΔWΔm=∑n=1NcpTi[(popi)γ−1Nγ−1], which approaches the isothermal compression work ΔWΔm=RTilog(popi) as N→∞. These observations lead to crucial differences between the design of DCAES and ACAES systems, highlighting that the conceptual similarity of simply swapping the combustion chamber with a TES is highly misleading. In reality, the switch between ACAES and DCAES means that the turbomachinery has to be completely reengineered. As a result of the second and third observations, DCAES has a greater number of compression stages with interstage cooling and fewer expansion stages with high inlet temperatures. For example, the Huntorf DCAES plant has 20 axial compression stages (pressure ratio 1.05) and 6 radial compression stages (pressure ratio 1.8), whereas the expansion is accomplished in just two stages.1Jafarizadeh H. Soltani M. Nathwani J. Assessment of the Huntorf compressed air energy storage plant performance under enhanced modifications.Energy Convers. Manag. 2020; 209: 112662https://doi.org/10.1016/j.enconman.2020.112662Crossref Scopus (24) Google Scholar Conversely, in ACAES, the charge-discharge coupling leads to a turnaround in the operational principle—the turbine inlet temperature should be lowered whereas the compressor outlet temperatures should be increased. These requirements go against the general trends for compressors and turbines because in compressors, losses tend to increase at higher temperatures and gas management becomes more challenging, whereas in gas turbines, higher inlet temperatures yield higher power extraction (gas turbine inlet temperatures above 1,200K are now common). Furthermore, the sliding pressure is another limiting factor in ACAES, resulting in lower system efficiency even with idealized components.2Barbour E.R. Pottie D.L. Eames P. Why is adiabatic compressed air energy storage yet to become a viable energy storage option?.iScience. 2021; 24: 102440https://doi.org/10.1016/j.isci.2021.102440Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar Although this is also an issue in DCAES, the system is less affected due to the larger number of compression stages, allowing for smaller individual pressure ratio variation, and fixed turbine inlet temperatures. This highlights the challenge that researchers face in designing compressors and turbines for ACAES. For compressors, highly adiabatic, high-pressure-ratio compressors are required. The predicted long-duration, high-power role for ACAES perhaps favors axial machines, but aerodynamic limits make simultaneous high-mass-flow and high-pressure-ratio machines very challenging to design, typically requiring transonic flow to provide the necessary mass flow and compression. It can be noted that, although some manufacturers do display compressor models able to reach HP ratios and flow rates, this does not mean that there is an “on-stock” model suitable for ACAES. Rather, it indicates a general design family that might be adapted, optimized, and designed specifically for an application. For the expansion, ACAES research might be able to learn from the development of more efficient lower temperature turbines (LTT) for other applications, particularly geothermal power production, pumped thermal energy storage (PTES), and low-grade heat-to-power and organic Rankine cycle (ORC) systems. As of the present time, LTT development specific to ACAES has been limited. Adapting steam turbines to ACAES applications is another possibility proposed in literature, as these generally operate at similar pressure levels but significantly lower temperatures than their gas equivalent.3Giovannelli A. Tamasi L. Salvini C. Performance analysis of industrial steam turbines used as air expander in Compressed Air Energy Storage (CAES) systems.Energy Rep. 2020; 6: 341-346https://doi.org/10.1016/j.egyr.2019.08.066Crossref Scopus (6) Google Scholar The compression-expansion coupling in ACAES also leads to major design challenges in the HEX. Equation 2 defines HEX effectiveness, ε—the ratio of actual heat transfer rate Q˙ against the theoretical maximum Q˙MAX—as a result of the generalized energy balance in a HEX (neglecting thermal losses to the environment).ε=Q˙Q˙MAX=Ch(Th,i−Th,o)CMIN(Th,i−Tc,i)=Cc(Tc,o−Tc,i)CMIN(Th,i−Tc,i)(Equation 2) Here, CMIN=MIN[m˙ccc,m˙hch] is the minimum heat capacity flowrate (the product of the fluid heat capacity and mass flow), m˙c and m˙h are the mass flow rates of the cold and hot fluids, respectively, and cc and ch are the respective heat capacities. Tc,i and Th,i are the cold and hot fluid inlet temperatures, whereas Tc,o and Th,o are the outlet temperatures. It is clear from Equation 2 that the TES can only achieve a temperature close to the compressor outlet temperature for a balanced HEX (i.e., m˙ccc=m˙hch) with high effectiveness. However, in classical design methodologies, high-HEX effectiveness ε>0.9 is achieved by using severely unbalanced conditions (CMIN/CMAX≪1). For instance, a compressor aftercooler utilizing water coolant might reach ε=0.9 with moderate surface area; however, normal coolant mass flow rates are around five times greater than the balanced condition, leading to a small coolant temperature change (20K, in comparison with 100K for the air). This imbalance allows for equipment compactness, with typical core volumes ≈0.5m3, reducing pumping losses and saving space. In contrast, if a 95% balanced flow condition is imposed, the HEX volume would increase by 6-fold and it would still only be capable of reaching effectiveness 0.75–0.8. In this case, however, reversibility is significantly superior.4Kays W.M. London A.L. Compact Heat Exchangers (Third edition). Krieger Publishing Company, 1998Google Scholar The first (unbalanced) HEX example is suitable for a DCAES plant, whereas for ACAES, the compression-expansion coupling requires the second approach. The primary design challenge for HEX in ACAES is finding the appropriate trade-off between effectiveness, reversibility, and cost. ACAES stores exergy both in (1) compressed air in the HP store and (2) thermal exergy in the TES. In the former, the pressure limits and volume define the energy storage capacity. The primary design criteria of thermomechanical integrity and air tightness over the system lifetime are also shared with conventional DCAES. These are well documented in technical literature for both aboveground and underground storage. However, in ACAES the storage temperature behavior must be properly integrated into the design process (e.g., Wang et al.,5Wang S. Zhang X. Yang L. Zhou Y. Wang J. Experimental study of compressed air energy storage system with thermal energy storage.Energy. 2016; 103: 182-191https://doi.org/10.1016/j.energy.2016.02.125Crossref Scopus (99) Google Scholar reports the importance of the storage temperature in the system performance and in Barbour et al.2Barbour E.R. Pottie D.L. Eames P. Why is adiabatic compressed air energy storage yet to become a viable energy storage option?.iScience. 2021; 24: 102440https://doi.org/10.1016/j.isci.2021.102440Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar losses associated with this temperature fluctuation are analytically addressed). Recent research has also conceptually outlined the thermodynamic benefits of isobaric air storage; however, the variable-volume HP air storage adds complexity and future research should explore this trade-off. The TES is a critical part of the ACAES system and must be seamlessly integrated with the HEX. Over half the energy generated in the two conventional DCAES plants currently in operation is due to heat from combustion, whereas in ACAES, there is no external energy input. The TES must store and maintain the heat extracted after compression until it is reintroduced prior to expansion. Thermal storage is also used in a huge range of industrial and domestic applications, and TES is a vast and parallel research field. Specific challenges to ACAES relate to the high temperatures (and potentially high pressures) involved and the long-duration and large-capacity requirements, which means that the TES must be simultaneously mechanically strong, efficient, and cheap. Moreover, the variable compressor discharge temperature (due to increasing pressure in isochoric storage) imposes the challenge of avoiding internal temperature mixing and the consequent losses. Aspects discussed in these sections are summarized in Figure 2. Numerous examples of theoretical and simulation studies on ACAES can be found in technical literature, i.e., with a predicted round trip typically in the range of 50%–75%.6Mucci S. Bischi A. Briola S. Baccioli A. Small-scale adiabatic compressed air energy storage: control strategy analysis via dynamic modelling.Energy Convers. Manag. 2021; 243: 114358https://doi.org/10.1016/j.enconman.2021.114358Crossref Scopus (5) Google Scholar,7Chen S. Arabkoohsar A. Zhu T. Nielsen M.P. Development of a micro-compressed air energy storage system model based on experiments.Energy. 2020; 197 (2020.117152)https://doi.org/10.1016/j.energy.2020.117152Crossref Scopus (28) Google Scholar Within these, the system individual subcomponents (i.e., compressors, heat exchangers, turbines) are generally based on “black-box” thermodynamic models, generating performance indicators from a given number of inputs without considering the internal component detailing. Although this approach is useful for conceptual studies and describing the general operating principles, it omits important equipment technical limitations and/or design challenges. This can result in unrealistic predictions of operating conditions and performance indicators. Papers that specify dynamic component performance also exist;8Sciacovelli A. Li Y. Chen H. Wu Y. Wang J. Garvey S. Ding Y. Dynamic simulation of adiabatic compressed air energy storage (A-CAES) plant with integrated thermal storage – link between components performance and plant performance.Appl. Energy. 2017; 185: 16-28https://doi.org/10.1016/j.apenergy.2016.10.058Crossref Scopus (139) Google Scholar however, these are still based on generalized models rather than specific custom-designed components. In a recent paper,2Barbour E.R. Pottie D.L. Eames P. Why is adiabatic compressed air energy storage yet to become a viable energy storage option?.iScience. 2021; 24: 102440https://doi.org/10.1016/j.isci.2021.102440Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar we derive idealized performance limits for isochoric ACAES systems, providing both a reference for finding the limiting performance and also specifying the target component operation to achieve this. The highlighted component operation illustrates that typical assumptions regarding component performance used in many papers can be misleading, given that they are based on examples from different applications with unique design requirements. Although some recent work does consider individual components designed and optimized specifically for ACAES, there remains a lack of publications covering heavily coupled components with consideration of internal design across the whole system. This is likely the reason for the disparity between simulated predictions and experimental measured performance at the current time. Most experimental studies are limited to studying ACAES components, rather than conducting full plant analysis. The TICC-500 pilot plant is a rare example of a published, comprehensive experimental study. The plant achieved a measured round-trip efficiency (RTE) of 22.6%, lower than its design efficiency of 41%.5Wang S. Zhang X. Yang L. Zhou Y. Wang J. Experimental study of compressed air energy storage system with thermal energy storage.Energy. 2016; 103: 182-191https://doi.org/10.1016/j.energy.2016.02.125Crossref Scopus (99) Google Scholar This performance decrease was attributed to the transient HP store pressure on the turbomachinery and TES, as well as lower than anticipated HEX effectiveness. Nevertheless, the work provides a strong foundation for future full-system prototypes. In 2016, ALACAES published results for a pilot plant located in Switzerland, reporting an estimated RTE between 63%–74%.9Geissbühler L. Becattini V. Zanganeh G. Zavattoni S. Barbato M. Haselbacher A. Steinfeld A. Pilot-scale demonstration of advanced adiabatic compressed air energy storage, Part 1: plant description and tests with sensible thermal-energy storage.J. Energy Storage. 2018; 17: 129-139https://doi.org/10.1016/j.est.2018.02.004Crossref Scopus (86) Google Scholar However, this efficiency was estimated by neglecting leakages, simulating an adiabatic compressor (with constant isentropic efficiency equal to 85%) and introducing an 90% efficient turbine, given that the compressor used was unsuitable for ACAES and no turbine was included. Disregarding the efficiency estimate, the main aim was to investigate the air tightness of the cavern and the performance of a packed-bed rock-filled TES. Both of these components performed satisfactorily under the reduced operational pressures. The vast majority of other experimental academic studies have undertaken small-scale tests on single components, typically at reduced pressures, yielding results that are difficult to scale-up. Furthermore, misleading performance metrics are often presented, such as peak-instantaneous efficiency (defined as a ratio of specific discharging power to specific charging power within a very narrow pressure range) rather than measured RTE (the ratio of total generated to consumed energy). A significant proportion of ACAES development has also occurred in commercial settings, often accompanied by optimistic, but opaque, performance claims. For example, the part-EU-funded project ADELE, originally intended to produce the world’s first large-scale ACAES plant, claimed to have proven that 70% efficiency is technically feasible,10Zunft S. et al.Electricity storage with adiabatic compressed air energy storage: results of the BMWi-project Adele-ING.in: International ETG Congress. 2017: 1-5Google Scholar however, no plant was ever built and design details were not published. The Canadian company Hydrostor operates two pilot plants and has recently announced it received substantive grants for R&D as well as two new plants. Performance metrics are not generally available, although information presented in Ebrahimi et al.11Ebrahimi M. Carriveau R. Ting D.S.-K. McGillis A. Conventional and advanced exergy analysis of a grid connected underwater compressed air energy storage facility.Appl. Energy. 2019; 242: 1198-1208https://doi.org/10.1016/j.apenergy.2019.03.135Crossref Scopus (34) Google Scholar suggests that the first plant (which is a proof of concept) has a very low efficiency. Notably, Hydrostor employs an isobaric ACAES design to mitigate the variable pressure in the HP store. Other well-funded commercial ventures with previously optimistic performance claims, such as Lightsail and SustainX, have failed to produce any pilot plants and have since been discontinued. Two notable plants are also mentioned in the extensive review by Wang et al.,12Wang J. Lu K. Ma L. Wang J. Dooner M. Miao S. Li J. Wang D. Overview of compressed air energy storage and Technology Development.Energies. 2017; 10: 991https://doi.org/10.3390/en10070991Crossref Scopus (125) Google Scholar claiming round-trip efficiencies of 55% and 60% for the 1.5 and 10 MW plants, respectively; however, the supporting references provided are defunct. Recent publications, major projects, and technology milestones are summarized in Figure 3. Despite the fact that the technology-readiness level (TRL) of DCAES is high, with two large-scale plants (we suggest TRL 8 due to lack of widespread adoption), the TRL of ACAES is low. At the current time, the TICC-500 experiment5Wang S. Zhang X. Yang L. Zhou Y. Wang J. Experimental study of compressed air energy storage system with thermal energy storage.Energy. 2016; 103: 182-191https://doi.org/10.1016/j.energy.2016.02.125Crossref Scopus (99) Google Scholar is the sole example of a full prototype plant published in mainstream technical literature. Furthermore , several commercial ventures developing ACAES have failed, despite very significant funding, leaving little in the way of documentation or lessons learned. Hence, we suggest that TRL 3–4 is appropriate. This puts ACAES at a similar TRL as PTES, another unproven but promising energy storage option, and arguably below liquid air energy storage (LAES), which has a demonstrator facility in the UK as well as a pilot plant at the University of Birmingham, UK. At first sight, this might be surprising because ACAES has a conceptually simpler layout than PTES and LAES and is conceptually similar to DCAES; however, as discussed, there are still major design challenges for a majority of components. As shown in Figure 3A, recent Hydrostor announcements of two new plants could improve the TRL of ACAES; however, this is dependent on their success or widespread dissemination of results and lessons learned. Figure 3B shows a comparison of the TRL across different technologies at the time of writing. Although conceptually well established, ACAES remains unproven at a viable performance level. However, there are a number of promising avenues for future research to explore. In particular, the adoption of a whole-system design approach, relating component performance to feasible designs (i.e., without black-box or generic component models) across the system, would be hugely valuable. These studies might then accurately reflect the achievable performance in pilot projects, which could subsequently be optimized for both performance and viability. In terms of component development, the simultaneous development of higher-outlet-temperature compressors and lower-inlet-temperature turbines is vital. This includes theoretical description, analytical performance prediction, CFD, and actual tests at different scales. Isobaric storage is also an interesting avenue to explore because of the potential to mitigate challenges associated with the transient pressure. Experimental research should focus on the design and control of the constituent components under real operational conditions. Full-plant prototypes in academic settings would hugely advance the state of knowledge surrounding ACAES. This research would be symbiotic with the development of other promising thermomechanical energy storage concepts, such as PTES, where the development and control of highly reversible turbomachinery is also a major challenge. Overall, given its potential as a clean and cost-effective electricity storage method for the future, we strongly recommend that funding bodies support further research in ACAES. Although the private sector is arguably better at raising significant capital, the short time horizons for which investors want return on investment combined with low TRL makes ACAES a difficult option. Furthermore, recent failures in the commercial sector have thus far yielded little in the way of lessons learned and contributed to uncertainty in the technology. Therefore, large-scale research in academic settings with a commitment to transparent documentation would be most worthwhile, and unverifiable performance claims should not influence funding awards. Finally, future academic articles should not blindly repeat performance claims without due scrutiny. If ACAES is to finally fulfill a protagonist role in the future energy market, research must first tackle key outstanding challenges that have been, so far, mainly overlooked. D.L.P. is supported by a scholarship provided by Loughborough University. The authors declare no competing interests.