Next Article in Journal
Exploiting Lubricant Formulation to Reduce Particle Emissions from Gas Powered Engines
Previous Article in Journal
Representing Carbon Dioxide Transport and Storage Network Investments within Power System Planning Models
Previous Article in Special Issue
Battery Storage Use in the Value Chain of Power Systems
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Energies
Volume 17
Issue 15
10.3390/en17153779
energies-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Long-Duration Energy Storage—A Literature Review on the Link between Variable Renewable Energy Penetration and Market Creation

by
Andreij Selänniemi
*,
Magnus Hellström
and
Margareta Björklund-Sänkiaho
Faculty of Science and Engineering, Åbo Akademi University, Domkyrkotorget 3, 20500 Turku, Finland
*
Author to whom correspondence should be addressed.
Energies 2024, 17(15), 3779; https://doi.org/10.3390/en17153779
Submission received: 24 May 2024 / Revised: 20 July 2024 / Accepted: 23 July 2024 / Published: 31 July 2024
(This article belongs to the Special Issue Review Papers in Energy Storage and Related Applications)
Browse Figures
Versions Notes

Abstract

:
The relationship between a region’s dependency on variable renewable energy (VRE) and the viability of long-duration energy storage (LDES) technologies is recognised through various electricity grid modelling efforts in the contemporary literature. Numerous studies state a specific VRE penetration level in total electricity generation as an indicator of the emergence of an LDES market. However, there is considerable variability across studies when comparing VRE penetration levels in conjunction with LDES technology utilisation, and significant diversity exists in electricity grid modelling approaches. This review aims to highlight these inconsistencies by offering an overview of disparate findings and dissecting the influencing variables. Sixteen parameters are identified from reviewed studies, complemented by an additional five recognised through in-depth analysis. This comprehensive examination not only sheds light on critical aspects overlooked in previous reviews, requiring further investigation, but also provides novel insights into the complexity of this correlation, elevating the understanding of LDES market creation by unravelling the factors that influence the technology adoption across various contexts. Furthermore, it provides clarity in LDES research terminology by rectifying ambiguous language in the existing literature. Altogether, seven databases were explored to produce a trustworthy foundation for the study.

1. Introduction

Electricity generated from variable renewable energy (VRE) sources has on a global scale been rapidly increasing [1], primarily due to the cost decreases in solar and wind power production and, more recently, in Li-ion battery storage capacity [2]. Additionally, policies and regulatory frameworks have been put in place to ensure the renewable energy trend continues to grow, enabling climate targets to be reached [3]. However, increased output from VRE sources challenges the current way of operating an electricity grid network due to the intermittent nature of solar and wind. As the VRE penetration increases, storage capacity in electrical networks becomes more important since it allows for a more optimised use of renewable energy through storage in times of excess supply. New grid-scale energy storage installations are predominantly configured with Li-ion battery technology due to the cost competitiveness in storage durations of up to around four hours. However, the challenge with the increasing VRE-dependency is that much longer storage durations are anticipated to be needed to enable a reliable and sustainable electricity supply [4,5]. Due to this, long-duration energy storage (LDES) has become a trending topic, as it enables a more optimised use of renewable energy [6]. Even so, LDES technologies are struggling with being deployed since longer durations are not yet needed [7].
The current literature exposes a broad variation in how LDES is defined, but this work follows Denholm et al.’s [8] recommendation and chooses the definition as ten hours of storage duration or longer. A central parameter of the commercial feasibility of LDES in energy systems is the VRE penetration level [%]. A key question is then at what penetration level LDES becomes commercially feasible (a tipping point here labelled market creation), to which the literature gives inconsistent answers. Several studies have made models where this link is visible, and clearly, many diverse variables influence the result. Nevertheless, there are still papers stating precise VRE penetration levels to define this linkage without providing details on fundamental considerations, such as an applicable region or country, or even describing the VRE percentage value as a general consensus, as seen in Ho and McNamara [7], Kaushik et al. [3] and Zingales [9]. Hence, a review is needed and to consider the best methodology for achieving the research objectives, striving to provide clarity on the range of values in the literature and the reasons behind existing differences, especially since the conditions and scale of VRE deployment necessary for transitioning from short- to long-duration energy storage are not comprehensively considered [5]. Current reviews predominantly focus on LDES role assessments, explored in Blanco and Faaij [10] and Schill [11], technology evaluations, investigated in Albertus et al. [12], and policy considerations, studied in Ratz et al. [13] and Ho and McNamara [7], leaving a significant gap in our understanding of how the increasing penetration of VRE sources influences the emergence and dynamics of LDES markets. To address this crucial knowledge void, this dedicated literature review aims to unravel the nuanced connections between VRE integration and the development of LDES markets, shedding light on essential aspects that previous reviews have not thoroughly examined.
The limitations of this research are tied to the chosen methodology. It is important to note that achieving comprehensive coverage of the relevant literature is challenging in any review, and this can lead to an incomplete collection of raw data that later affects the analysis. Additionally, the scope of the research is restricted by specific inclusion criteria for practical purposes, potentially overlooking articles beyond this scope that could be applicable to the examined topic. Furthermore, the inherent varying quality of studies adds complexity when comparing results. Nevertheless, these limitations were foreseen and mitigated by diligently selecting and scrutinising high-quality databases and later complementing the literature found by search strings, increasing confidence in the accuracy of the findings.
The structure of this article is organised in a logical sequence, beginning with a comprehensive examination of the methodology employed in the review. This methodology section serves as the foundation for the subsequent Section 3, where an overview is provided of the presented findings in the current literature concerning VRE penetration and LDES utilisation. Following this, the article proceeds to examine the key factors contributing to these findings, whereafter the article concludes by synthesizing the research outcomes and providing additional clarity on the significance of the study’s results, fostering a more comprehensive understanding of the subject.
The primary outcomes and contributions are twofold. Firstly, this article provides an extensive overview of the existing literature, highlighting the notable divergences regarding the relationship between VRE penetration and LDES adoption. Secondly, the research delves into the critical variables that underlie these disparities, clarifying the factors that influence LDES adoption within different contexts. This in-depth analysis not only enriches our understanding of LDES market creation but also identifies areas in the field that have been less explored and merit further investigation. In our assessment, this review is essential to bridge the gaps that remain between prior publications in the field.

2. Materials and Methods

This literature review analyses existing studies to find established connections between VRE penetration and LDES market creation. In this context, VRE penetration refers to the proportion of total electricity generation attributed to variable renewable energy sources. Market creation denotes the point at which LDES technologies become competitive and are widely adopted. Therefore, this review’s primary interest is various papers focusing on modelling a hypothetical electricity grid in a future scenario with plenty of renewable energy utilisation. Papers not conducting any modelling are also included to establish a better understanding of the current state and research objectives in the LDES literature.
This review is conducted solely and manually by the authors without any automation tools besides standard exclusion options in search efforts. Study characteristics used as criteria for eligibility (i.e., “inclusion criteria”) are the following:
  • Published between 2015 and 2022;
  • English language;
  • LDES mentioned in the abstract;
  • Wide perspective on the LDES subject;
  • Main focus on the power sector of a larger system;
  • Evident link between VRE penetration and initial LDES technology deployment;
  • Peer-reviewed;
  • Full text available online;
  • Article as document type;
  • Available in one or several of the targeted databases.
These requirements create a concentrated result of papers to be analysed. The first criterion (1st) was chosen to avoid outdated information influencing the review’s findings—as price declines and market movements in the energy storage field change quite dramatically in only a few years. The second (2nd) and third (3rd) criteria are picked to ensure all papers are interpreted correctly and that the LDES subject is a focal point of the reviewed research. The fourth (4th) requirement is in place to establish a collection of studies that do not investigate a detailed process or specific technology but investigate the broader topic of LDES instead, as such a collection fits the purpose of this literature review better. The fifth (5th) and sixth (6th) standards were chosen to align the assessed publications with the interest of this study, and the seventh (7th), eight (8th), ninth (9th), and tenth (10th) requirements are used to restrict the review to high-quality, easily accessible, and up-to-date information. Criteria (1), (2), (3), (7), (8), (9), and (10) are used as exclusion options in the search, while criteria (4), (5), and (6) are added afterwards to the results of the database scan to further extract the relevant literature.
Altogether, seven databases were searched on the 7 May 2024 with criteria (1), (2), (3), (7), (8), (9), and (10), with noted results as per Table 1.
Based on these 1057 search results, another sorting was done to remove duplicate articles, whereafter criteria four (4) and five (5) were applied, leaving the intended analysis with 84 papers to be studied. As the literature review intends to find established connections between the VRE penetration rate and initial LDES technology usage, these 84 papers were further reduced by adding requirement six (6): inspection of the content to find indications or statements of such a connection. This inspection was conducted by carefully scrutinising the papers, reading through the text and examining figures, tables, and possible Supplementary Materials. Of these 84 articles, a link between the two factors can be found in 26. As this entailed a more detailed analysis, seven more technical studies were discovered containing applicable information to the review at hand. Altogether, these 33 papers (enumerated in Supplementary Materials—File S1) create the foundation of this literature review. Notably, the additional seven studies included do not follow all the original ten criteria since these studies were discovered through the process of snowballing, as visible below in Figure 1.
All 33 sources were examined using the same framework to extract the needed data: each reference was categorised and summarised in Microsoft Excel based on chosen assumptions and a modelling approach. The link between the VRE penetration rate and LDES technology usage in the explored set of papers is not the primary research target, but the connection can still be identified. In some cases, e.g., Du et al. [1], the information is already evaluated in the text and can easily be extracted. In others, e.g., Bertsch et al. [14], a more thorough examination is needed where the data is collected from the article and then aligned to be comparable with the rest of the results in the overview, meaning that for each source established, there is a specific VRE penetration percentage point at which LDES technologies are first utilised or mentioned to be needed.
The current literature acknowledges that flexible resources will be required to achieve a reliable and zero-carbon power system [5]. Two references are therefore included in this review that model LDES in combination with 100% renewable energy, demonstrating the need for LDES and its suitability in providing this flexibility [15,16]. However, since this study is devoted to the subject of LDES market creation, focus lies on reviewing research where the progressive development of such a power system is visible. No study indicated that LDES is only needed at a full-scale renewable energy configuration; thus, it is assumed these technologies will be needed before that level of development.
Regarding the observed variability in the reviewed literature’s findings, Table 2 reveals 16 variables that influence the link between VRE penetration and LDES adoption. These were identified through an analysis using three distinct inclusion criteria, where the first involved identifying variables that are recognised as key drivers for LDES deployment. The second criterion expanded the scope to encompass other factors that have been shown to influence the required flexibility of a modelled system, while the third was centred on elements that either constrain or frame the outcomes of a study. This set of inclusion criteria comprehensively encompasses three key aspects that influence the researched topic, thereby curating a collection of highly relevant variables (some of which apply to multiple criteria).
The existing research diverges in how these 16 variables are taken into consideration, as visualised in Table 3. Next, in the forthcoming Section 3, the variables are further explained and explored based on observations and correlations in the reviewed literature.

3. Results and Discussion

Figure 2 is a visualisation of the result from Table 3, showing that 80% VRE penetration is the most referenced value for the tipping point of LDES market creation. However, this review reveals a significant inconsistency in how the literature portrays LDES utilisation in relation to VRE penetration, with values for VRE penetration ranging from 20% to 100%. A linear trendline is included (y = 0.006x + 0.0079) to indicate that the reference amount grows with growing VRE penetration percentage values.
Considering this outcome, it is obvious that it is not possible to conclude a general VRE penetration level that would indicate when LDES reaches commercial feasibility, although this, as mentioned, was done in previous studies like Ho and McNamara [7], Kaushik et al. [3] and Zingales [9]. The variance in results stems from the diverging approaches taken with the 16 variables as visualised in Table 3, and these variables will be further assessed and discussed next.

3.1. Modelling Tool and Purpose

The tool for investigating one’s parameters creates the framework for achieved results. Each tool has its limitations and benefits, ultimately shaping the research conducted [35]. For instance, Zhang et al. [25], utilise the PLEXOS tool to model four separate LDES types but restrict the analysis to use a pumped storage object for all to mirror the capabilities, leading to the only separating factor to be the round-trip efficiency—which, in many studies, has been considered a secondary characteristic for LDES technologies [4,10,24]. Table 4 shows that many different modelling instruments exist, as nearly all included sources have used different ones. A tool should also be chosen based on the purpose behind one’s research [35], further influencing the results in the compared studies. For example, Leonard et al. [18] are interested in simply modelling a power system that would function reliably based on 25–30% variable renewable energy sources, while de Sisternes et al. [6] are interested in not only reliability but cost optimisation. An energy system model, which includes a cost parameter compared to one that ignores it, will have a different conclusion. Table 4 below presents a comprehensive enumeration of the various tools utilised, and an encapsulation of the underlying objectives guiding the modelling endeavours undertaken in the reviewed literature.
Key focus areas in the modelling approach for the reviewed literature appear to be cost optimisation, role evaluation, and functionality exploration for LDES technologies and systems.

3.2. Model Horizon and Approach

The time horizon of the modelling attempt has also shown to be crucial when demonstrating the value of LDES [15]. The literature contains discrepancies on the chosen model horizon, but as Table 3 shows, many sources use a period of one year. Nevertheless, some studies have stated that a year is insufficient to display the advantages of LDES [15,34]. If this is accurate, a significant part of the examined research might establish a need for LDES at lower VRE penetrations if these diverse systems are modelled over longer periods. Of course, this also depends on the model’s design. Most of the reviewed literature uses hourly resolution to study assumptions over the designated period, an approach that de Sisternes et al. [6] also take. This paper aims to model a least-cost system in ERCOT (USA) in 2035, but instead of modelling the entire year, only four weeks of data are analysed to mirror the whole term.
Conversely, in Dowling et al. [15], statistics between 1980 and 2018 are examined in six-year periods to include as many infrequent weather events as possible to get a clear overview of the inter-annual variability in output from VRE resources. However, the difference is that de Sisternes et al. [6] are trying to model a future scenario based on estimates, whereas Dowling et al. [15] are looking at a past time interval with actual data and fitting a hypothetical power generation system into that. This creates a less complicated model as assumptions are fewer, enabling a longer horizon. Weitemeyer et al. [30] is another source with the same approach, looking at data between 2000 and 2007 to check demand profiles and VRE electricity generation potential—to establish how much energy storage would have been needed with the proposed hypothetical power system. Many elements in the modelling approach vary between studies and influence the results, but it can be concluded that the modelling horizon is a key feature to consider for LDES. A longer horizon will reveal larger seasonal variations in weather, exposing the need and illustrating the benefits of LDES for complete system optimisation at various penetrations of VRE.

3.3. Geographical Market

The scoped region (its size and circumstances) naturally has a significant impact on the outcome of one’s model, e.g., as in Shaner et al. [29], where the wind resource peaks in spring, while the electricity demand is at its minimum. This match or mismatch between electricity generation and demand establishes the opportunity for energy storage. Depending on the chosen location, other flexibility options might directly influence the need for LDES. Some countries have abundant access to solar energy, while others are richer in wind. One nation may be primarily geographically flat, while another has big mountains and height differences. All these factors influence the feasibility of how and when renewable electricity can be generated. Moreover, these circumstances affect the feasibility of various storage technologies, e.g., compressed air energy storage (CAES) is an economically unattractive choice for excess wind power in the Danish electricity system unless it can defer investments in generation capacity. However, in Germany, CAES can be economical under certain wind penetration levels [36]. If the area is large enough, some of the variability can be counteracted with transmission lines, as weather conditions often differ among regions, but with a very restricted geographical scope, this might not be the case. Thus, LDES typically has a larger influence on smaller modelled systems [15], and generally, the bigger the modelled area, the smaller the optimal storage capacity [11].
However, a larger baseload supply will also increase the chances of surplus generation from VRE resources, elevating the value of energy storage and the business case of LDES. Although, with more immense geographical coverage, the time resolution or spatial granularity must be smaller in the modelling phase, which might lead to oversimplifying the whole system [10]. In the reviewed literature, the modelled area ranges between 22,145 km2 (Israel) [37] and 510,072,000 km2 (global) [38]—over a 23,000% difference. Thus, predictions and assumptions between these studies diverging is inevitable. Hence, proceeding with caution is recommended if the previous literature is used as a reference when specifying the link between VRE penetration and LDES market creation. For example, Albertus et al. [12] state that LDES will be needed at a VRE penetration of 70–90%, using two separate studies as references that apply different assumptions, where one conducts modelling of the entire USA but the other only specifically to the ERCOT region.

3.4. Assumed Electricity Demand

Another consideration when identifying reasons for varying results in the current literature is if and how fast the electricity consumption is estimated to grow. An example would be a scenario where a community is modelled to go from situation A to situation B, where two models with the same modelling horizon and circumstances are compared, but situation B differs in the amount of needed electricity. In the scenario with the higher electricity demand, more generation is built during the same period, which will likely involve VRE assets. This increase in variable electricity generation, combined with other assumptions made in the models, may lead to deployment of LDES technologies at a lower level of VRE penetration. Such assumptions could, e.g., be related to emission requirements, the retirement of conventional generation, or the feasibility of other competitive balancing technologies (e.g., as in Armstrong et al. [39], where natural gas generation without carbon capture and storage [CCS] is discarded due to a 0 gCO2/kWh requirement). Therefore, it is possible that the estimated electricity demand only indirectly influences the LDES market creation but is still a variable in play.
Compared to actual grid operation, one drawback with grid modelling is that the electricity demand is estimated based on average historical data and is always known. Unforeseen events may be overlooked, which can fabricate results with little uncertainty compared to a real-life scenario, creating small backup generation reserves and security of supply—a technology fit for LDES [4]. The future anticipated electricity demand in modelling efforts is also very region dependent. Consequently, few studies can be compared in Table 3, especially considering that not all studies make an assumption on the growth. Nevertheless, Table 5 is an overview of the literature that can be analysed. These demand figures are incompatible with the VRE % visible in Table 3, as the Table 5 data are based on 2050 for comparison reasons, besides for source [24], which is based on 2030. Sources [22,25,32] all make strikingly similar assumptions, while sources [5,24] assume corresponding values but with a 20-year difference. Sources [2,14] also differ by around 18%, a considerable deviation.

3.5. Assumed Variable Renewable Energy Penetration (VRE %) Level

Regarding the market establishment of LDES, the level of VRE instalments is crucial, as the business case of LDES technologies has its foundation in the intermittency of these renewable sources [40]. If the deployment scale of solar and wind power assets is small compared to the full fleet in a particular system, the use case of LDES diminishes as the variability in output from renewable sources is a smaller concern. Thus, the assumed VRE penetration rate in modelling efforts will largely determine the competitiveness of these technologies.
What must be mentioned about the noted VRE percentages for each source in Table 3 is that the accuracy varies on how well this percentage describes when the market is created for LDES in each model, because the initial investments in these resources are not visible or declared in all studies. For example, in Belderbos et al. [31], an initial VRE penetration of 50% is first assumed, where short-term battery storage is preferred in the model optimisation, after which 80% VRE penetration is examined to show that power to gas (a typical LDES alternative) becomes the preferred storage instead. However, this could still mean that LDES technologies become competitive in an earlier phase, but this cannot be confirmed as there is no visibility into the system development between 50% and 80% VRE penetration. Some sources also state an interval of when LDES becomes feasible, e.g., Ho and McNamara [7], which mentions between 70% and 80% VRE penetration, but as this literature review focuses on market creation, the lower end is chosen for Table 3.

3.6. VRE Mix

The amount and balance between solar and wind power utilisation typically vary among models, although the same VRE penetration rate might be assumed, which is visible in Table 3. A major reason behind this dissimilarity is that the VRE portfolio tends to be optimised based on what is most economically feasible for a region; if regions vary between models, so will the VRE balance. Yet, as Table 5 identifies, the weight between the sources may still differ, although the specified location is the same due to other diverging assumptions made in the modelling stage, e.g., if wind power is considered one asset class or if onshore and offshore are modelled separately with varying benefits, as done by Schill and Zerrahn [28]. This is not considered in all reviewed studies, which is an intriguing finding since offshore wind power has been acknowledged in the literature as having the potential to provide more firm generation [28].
Table 3 shows that the larger share of the total VRE portfolio often seems to be held by wind. This matters from an LDES market-creation perspective because, by assessing the different profiles of generation between wind and solar, it has been declared that a region with a more wind-heavy portfolio might create larger needs for LDES due to more extensive seasonal variations in output [5,10,15,23]. Lund et al. [36] also found that wind speeds tend to be higher at night, necessitating longer energy storage durations for an optimal system. As the electricity generation profiles between solar and wind power differ, they may even complement each other [10,12,15,18,29,30], which also influences the potential demand for LDES, as an optimised VRE portfolio will increase grid stability.

3.7. Weather Conditions

Weather conditions largely influence a system based on energy storage in combination with vast installations of VRE, which is why assumptions on this matter can heavily impact results in modelling efforts [29]. However, the influence of assumed weather conditions largely depends on the entire system configuration and the availability of flexible power.
In the reviewed literature, weather conditions vary based on the chosen region and the approach taken for the data collection. For example, in de Sisternes et al. [6] and Schill and Zerrahn [28], only one year of weather data are collected, unlike Weitemeyer et al. [30], Ziegler et al. [34], Shaner et al. [29], and Dowling et al. [15], where 8, 20, 36 and 39 years are considered, respectively. Several years of data are essential to account for seasonal variations, as this directly affects the amount of storage needed and its required duration [15,41]. Moreover, climate change significantly impacts weather conditions. Despite ongoing debates within and outside academia, LDES modelling often inadequately accounts for these changing environmental conditions, as detailed in Section 3.17.1.

3.8. Consideration of Past Installations

This aspect is integral to accurately estimating the competitiveness and future use of LDES technologies. Still, as Table 3 shows, many studies neglect it. In practice, the timing of wide-scale LDES deployment will largely be affected by the competitive landscape formed by other technologies suitable to provide the needed balancing power [34]. In today’s situation we also need flexibility to enable our electricity supply to meet demand [36]. This flexibility, which is currently provided by assets such as natural gas power plants with an expected operating life of 25–30 years [42], will increase in importance as the influence from VRE sources increases [9]. All coming and existing short-duration li-ion battery deployments in recent years will also impact the competitiveness of LDES, especially considering the derating possibilities to enable longer durations [8]. These installations have the first-mover advantage with ancillary services and capacity payments, which may be largely covered. LDES technologies might need to find new ways of demonstrating value and generating income, which is why revenue streams are difficult to predict. Without substantial cost reductions, this remains a major challenge for LDES to overcome [17].
Therefore, evaluating a future scenario and market potential for LDES unaffected by the past is unrealistic. For instance, Dowling et al. [15] noticed significant decreases in energy storage deployments in a least-cost optimisation model when natural gas power plants were included in the system. Studies that assume a range of resources beyond VRE typically result in cheaper decarbonisation, even when continued reductions in wind, solar, and storage costs are assumed. These studies can also yield physically smaller grid systems because they avoid overbuilding VRE and transmission to meet seasonal demand. In these studies, energy storage has a role, but optimal durations are smaller—in the 4-to-16-h range [13]. De Sisternes et al. [6] also found that the most cost-effective approach would be to include nuclear power; the results from this paper indicate that energy storage may be essential to enable climate mitigation strategies dependent exclusively on very high shares of wind or solar energy, but storage is not a requirement if a more diverse mix of flexible, low-carbon power sources is considered. Other studies have achieved similar results [10], indicating that firm capacity from conventional generation could play a vital role in the energy transition to lower total power system costs, especially considering the prospects of carbon capture and storage (CCS) [43]. Some authors have highlighted that the availability of these zero-carbon firm technologies could diminish the need for LDES, but most of these studies fail to correctly model LDES due to oversimplification [41]. Bertsch et al. [14] even concluded that CCS might enable coal-fired power plants to co-exist with strict emission requirements.
However, CCS is still an immature technology with the potential of being a high-cost, high-risk solution on economic, environmental, and social grounds [43]. One reason is that CO2 is not the only harmful emission associated with burning fossil fuels, and the technology does not address sulphur oxide, nitrogen oxide, or heavy metal emissions [2]. Guerra et al. [5] also found that a high carbon-free or renewable energy mix could be cheaper than relying on CCS and natural gas. Nevertheless, research that only considers renewable energy assets in combination with short- and long-duration energy storage without accounting for the location’s current ways of generating electricity may arrive at questionable results, and the work adds little value to actual regional development. Other practical aspects essential for realistic modelling include safety considerations, as well as reliability and availability constraints, which are further detailed in Section 3.17.3 and Section 3.17.5.

3.9. Applied Emission Requirements

The ongoing decarbonisation of the energy sector, including stricter emission requirements, is one of the main drivers for the market creation of LDES [4]. Generally, the more stringent permits, the faster the market opens, as fossil fuels have increased difficulty competing in VRE balancing [6,43]. This effect is also shown in Childs et al. [24], where LDES needs increased substantially when a more restrictive greenhouse gas target was adopted. However, in Gaete-Morales et al. [20], a carbon tax is implemented while coal generation is still doubled, showing that varying assumptions and model circumstances will naturally influence the results achieved for when LDES will be necessary to provide grid stability and reliability. Modelling efforts are also typically cost-optimised, but practical examples of issues this approach can lead to can be found in the Dutch and the Irish power system. Here, energy storage may actually increase the overall CO2 emission levels since the technology allows storing power from cheap coal plants to substitute expensive gas during peak demand [36].
Table 3 shows that not all reviewed sources consider emission requirements, and when considered, assumptions and approaches differ, as Table 6 outlines. Sources [4,14,20,26,28] place a price on carbon emissions; through that, the model optimises the system cost by proposing feasible technology deployments. Conversely, sources [6,15,24,27,31] instead restrain the total system footprint through emission limits.
Clearly, from Table 6, if CO2 emissions are considered, stricter emission limitations in the future also tend to be assumed, which could contribute to why LDES typically becomes feasible at later stages of the model development (as Figure 2 shows). As Bettoli et al. [4] noted, a slower transition to net-zero emissions in the power sector will directly influence the amount of deployed LDES; de Sisternes et al. [6] even state that energy storage is only strictly necessary to meet tight emissions limits in the absence of other flexible dispatchable zero-carbon generation technologies. Dowling et al. [15] also demonstrate this competitive landscape where LDES is deployed fully in a 100% carbon-free power system, but as the limit is lowered to 95%, natural gas is also utilised.

3.10. Available Balancing Technologies (LDES Type)

The literature has established that a combination of balancing technologies with varying durations and characteristics will create the most optimised system [4,5,30]. Moreover, certain technologies, such as flow batteries, are seen as particularly promising for stationary energy storage [44]. Therefore, assumptions in the modelling stage on what LDES technologies are available will impact when the LDES segment becomes competitive when weighted against the VRE penetration rate, for example as in Belderbos et al. [31], where only power to gas (P2G) is available as an LDES resource, compared to Schill and Zerrahn [28], where also pumped hydro energy storage (PHES) is considered. PHES may be more competitive in shorter durations than P2G, possibly leading to LDES utilisation at lower rates of VRE penetration (depending on other available flexibility options and model assumptions). This is visualised in Du et al. [1], where an overview was created of typical energy storage technologies and their suitable durations (Figure 3).
Regarding the reviewed literature, Table 7 summarises LDES technologies considered in the analysed models.
Besides these LDES resources, other flexible technologies are considered in many models, e.g., in Gaete-Morales et al. [20] and Guerra et al. [22], where flexible natural gas assets are included, ultimately influencing the LDES deployment. This impact is also observed in Bertsch et al. [14], Hunter et al. [32], and Nikolakakis et al. [27], where NG + CCS is utilised and mentioned in Table 7 (although not considered an LDES technology) since the comparable use case with long-duration energy storage technologies is so evident in these papers. Li-ion is also considered for durations above 12 h in Hunter et al. [32], which is why it is listed. Another note on the reviewed literature is that there seem to be two main ways of approaching LDES technology modelling (separated between sources [15,24] in Table 7). One way considers a specific technology, while the other uses a generic model when estimating values on costs and efficiencies.

3.11. Estimated Costs and Efficiencies

Cost reductions of LDES technologies will drive their potential large-scale utilisation [7,34]. For example, Bettoli et al. [4] found that LDES costs would have to decline by 60% for the segment to become competitive. Round-trip efficiency was discovered not to be the most crucial characteristic for these types of technologies as they are reserves in periods of low VRE output, but according to Guerra et al. [5], the cycling behaviour is still closely tied to the efficiency, which tends to be lower for LDES technologies compared to short duration alternatives [31]. Hunter et al. [32] also found that the utilisation rate of LDES technologies increases when the round-trip efficiency increases, indicating that forming an LDES market could be accelerated through efficiency improvements.
Consequently, assumptions on these matters directly influence the timing and benefits of LDES installations in generated models. These assumptions are the primary factor determining whether the technologies are cost-competitive with alternatives in optimisation scenarios, thereby impacting the entire model, as evident in Section 3.8 on considerations of other installations and Section 3.16 on curtailment. As a result, cost and efficiency estimations are complex, arising from a series of other assumptions on elements such as learning rates [4]. Additionally, real-world factors like supply chains and political relations (discussed in Section 3.17.2 and Section 3.17.4) further complicate the accuracy of modelling. Table 7 gives an overview of applied estimates, where technology-specific assumptions tend to be similar for the same period, with some exceptions (e.g., CAES efficiency of 84% in source [2] but 60% in source [45]). Data transparency varies across the reviewed literature, resulting in certain fields in the table containing additional annotations within brackets to clarify the indicated values.
Compared to shorter duration energy storage (e.g., Li-ion batteries), LDES assets tend to have higher power capacity costs (€/kW) but lower energy capacity costs (€/kWh) since these technologies are optimised to economically provide long-term energy storage [5]. In fact, Denholm et al. [23] found that the closer you reach zero in costs associated with duration (energy capacity), the faster the market is opened for the technology. It has been declared that LDES should reach values of 5–35 €/kWh to compete with other alternatives [12], and this is also reflected in Table 7’s estimations. As these technologies are cycled infrequently, however, it is crucial for the feasibility that the added value provided through longer durations is being incentivised and acknowledged [4].

3.12. LDES Value and Business Synergies

Besides the needed cost declines of LDES technologies, a major challenge is receiving regulatory and policy support to kickstart and initiate the market [17]. For LDES to become a sound investment, market mechanisms would have to be restructured [22], e.g., in the U.S., several markets give assets with 4 h of energy storage full capacity credits—leading to a lack of incentives to invest in longer durations [45]. LDES would have to be recognised for its added value from a full system optimisation perspective, as in the Childs et al. [24] modelling, where only storage technologies with durations of 10 h or more receive full capacity credit. This distinction is important since the benefit of LDES is chiefly driven by its capacity value (essentially its ability to replace fossil-based peaking generators) rather than by operational value (load-shifting benefits) [22]. This connection to more conventional generation is also seen in Zhang et al. [25] since most of the LDES asset value comes from avoided startup and shutdown costs as well as avoided fuel costs for thermal generating units.
Consequently, how various sources in the literature consider LDES value and business synergies and what grid services that are assumed to be provided will affect the advantages of the technologies compared to other alternatives. A typical approach, like Du et al. [1] and Ziegler et al. [34]’s, is to simply model according to an assumed cost per kW and kWh for various technologies and then look at LCOE values. However, in Guerra et al. [22], the installation’s profitability is also considered through a separate metric to examine not only the total cost but the total system value. Hunter et al. [32], although not covered in the baseline case, include a sensitivity analysis with co-producing hydrogen to cut costs—another good example.
In Albertus et al. [12], only two revenue streams are considered for an LDES asset; one is electricity price arbitrage, which Blanco and Faaij [10] deemed unfeasible for LDES resources with durations above 24 h. Therefore, the LDES business case differs between the reviewed sources, which is logical since the VRE penetration level and asset portfolio influence the value LDES technologies can provide [11]. An example is load-levelling services designed for 1–10 h of duration since this service is typically more cherished in wind-heavy systems as wind speeds tend to be higher at night, leading to the required energy storage duration being longer [36]. Thus, from a modelling perspective, the overall benefit LDES can provide is closely linked to the assumed duration.

3.13. Storage Duration

Table 3 shows that the reviewed literature considers a range of durations. However, what is common for all is that use cases above 10 h are studied. The duration matters when exploring the link between VRE penetration and LDES market creation because different grid services relate to different energy storage durations [36], as visualised in Table 8.
For example, if one source is considering a duration of 10 h and another of 100 h (for the same technology), the timing will differ on when competitiveness occurs, which is also closely tied to the previous sections on available balancing technologies, LDES value considerations, and cost assumptions. However, the duration as such is still kept a separate variable as there are sources in the reviewed literature where neither of these previously mentioned aspects are considered, but a duration is yet determined (as Table 3 shows). With a long enough duration, energy storage assets can also compete for investments otherwise planned for transmission infrastructure [13].

3.14. Assumed Interconnections

The foundation of the LDES value proposition is to enable carbon-free variable renewable energy balancing over extended periods to secure electricity supply in periods of low sun and wind power generation [13]. In theory, this same contribution can be provided by transmission lines through interconnections to other countries and regions, allowing renewable energy-based electricity imports from locations with different weather circumstances when challenges arise with levelling supply and demand [46]. Thus, vastly different results can be achieved for required LDES capacities when considering a closed or open system for a modelled area.
Naturally, if interconnections are assumed, it will affect the deployment of LDES technologies and the need for other balancing power [2,11,47]. In fact, Blanco and Faaij [10] found that global energy storage needs decrease by 3× when the transmission is optimal between countries. However, many new interconnections would have to be established for optimal conditions, which may be more expensive than energy storage deployments, as a separate study found that storage systems can be up to 40% cheaper than expanding transmission infrastructure [48]. Following that logic, Bettoli et al. [4] noted that the U.S. had the greatest need for LDES systems among the modelled locations in the report, mainly due to the limited transmission connections across the country.
Interconnections have practical challenges when combined with strict emission targets, as electricity imports may be unspecified [24]. Generally, a combination of LDES and interconnections will create the most cost-optimised system [10]. However, this again depends on the other variables mentioned earlier. For example, Sepulveda et al. [43] found that interconnections only moderately impact the full system cost when firm capacity is available. Energy storage may also require agreement from fewer decision-makers compared to transmission infrastructure expansion and can be easier to implement, as well as achieve greater carbon emission reductions than using backup generation, e.g., natural gas turbines or engines [34]. Still, if both interconnections and LDES are utilised, the region’s electricity generation system would not have to be designed for 100% availability and self-sufficiency, which could be more cost-effective [34]. Conversely, Shaner et al. [29] questioned the feasibility of pursuing both options simultaneously.
The size of the modelled market is also closely linked to this subject, as an optimised transmission network does not necessarily have to mean interconnections to other countries if the area is big enough (e.g., as in Dowling et al. [15], where the full U.S. power system was examined). A developed transmission network, in conclusion, provides flexibility, which is needed to secure supply.

3.15. Demand Side Management

This section combines sector coupling, storage-to-storage operation, and demand-side applications, which are additional ways to add flexibility to support the wider grid challenge of matching supply with demand. For example, Blanco and Faaij [10] discovered that when the wider energy system is reviewed, and other alternatives besides power are examined, it seems there are options more attractive than storage. Schill [11] also noted the possibility of sector coupling to substantially decrease energy storage needs, causing an altered business case of LDES when considered. Storage-to-storage operation varies in utilisation between the reviewed sources, as it, e.g., is prohibited in Zhang et al. [25] while utilised in Bogdanov and Breyer [33]. Such mechanisms may enable optimised short-duration battery charging and more regular operation of the LDES system [2].
Since demand-side applications change peak load and momentarily required generation capacity, DSM can be seen as a source of flexibility [3]. Hence, if considered, it will influence the modelled requirements for other balancing power such as LDES. Strbac et al. [49] even observed that a flexible demand of only 20% of the peak demand could reduce energy storage benefits by almost 80%. However, this is most applicable in lower VRE penetration levels, as energy storage is then mainly necessary to enable a more effective use of renewable energy, not to secure the electricity supply. Thus, DSM (and sector coupling) are most functional for short-term balancing, not for empowering long-term grid reliability [11]. Additionally, the marginal value of each increment of demand flexibility declines as VRE and energy storage costs fall [43], and energy storage may allow greater quantities of electricity to be time-shifted when needed [34]. In effect, greater demand flexibility reduces (but does not eliminate) the need for excess installed VRE and storage capacity to reliably meet demand [43].
Demand-side applications in this review also include assumptions in the reviewed literature on utilising electric vehicle (EV) batteries for demand-side mechanisms and other residential energy storage or electricity generation setups for increased self-sufficiency, as such systems can help balance VRE [11]. For instance, Child et al. [2] noted a 17% decrease in electricity grid consumption when prosumers were considered. Conversely, Shaner et al. [29] concluded that the ability of vehicle-based storage to buffer gaps between supply and demand is limited, and Lund et al. [36] report that while EV services to the grid may decrease the need for overbuilding VRE (with connected risk of curtailed energy), other balancing mechanisms will still be needed.

3.16. Curtailment Possibilities

As Table 3 shows, VRE curtailment (deliberate reduction of potential output) is allowed or prohibited in the reviewed sources, creating dissimilarities in the literature. While unfavourable from an energy utilisation perspective, it can still be beneficial from a system cost optimisation point-of-view. This has been shown in multiple studies, where allowing a small percentage of energy to be curtailed creates a lower electricity price than developing a system with full storage for all generated electricity [1,5,11]. According to Blanco and Faaij [10], curtailment is the best option below 30% VRE penetration since the number of hours where there is a surplus is insufficient to justify an investment in any asset (of course, this conclusion is subject to the paper’s assumptions).
Conversely, if no energy is allowed to be curtailed, it will create more opportunities for energy storage (LDES included), although this might not be cost-optimised [10,11]. Therefore, the curtailment level allowed in grid model developments will influence the deployment of LDES technologies. Still, curtailment could be avoided if overall power system flexibility was increased, e.g., by having fewer must-run base-load power plants in the system or by using load-shifting [36]. On the contrary, Guerra et al. [5] argue that power curtailment remains a cost-effective flexibility option for integrating VRE sources despite large-scale installations of different storage technologies. Subsequently, when VRE shares are high and LDES becomes feasible, curtailment decreases.

3.17. Additional Findings

This section introduces five variables relevant to the subject matter, but they have been excluded from Table 3 since they were determined to have no impact on the analysed results or were unaddressed in the reviewed literature. A separate segment is included about renewable phrasing, aiming to establish a uniform vocabulary for future LDES research.

3.17.1. Climate Change

LDES technologies are considered to enable a fossil-fuel-minimised or completely fossil-fuel-free energy sector without jeopardising a reliable supply of electricity by storing energy from VRE sources for extended periods [4]. Reduced fossil fuel usage relates to the ongoing global trend of minimising our carbon footprint to mitigate climate change since it is otherwise expected that the number of extreme weather events will increase, and temperatures will rise [4]. This would not only affect electricity demand but the spatial and temporal distribution of solar and wind power [29]. Even so, not a single reviewed source takes this into account in weather considerations, as past data are used to forecast the future. This might be a more logical approach compared to making new assumptions, but it still involves the risk of having too conservative estimates. Some studies conduct sensitivity analyses over a series of scenarios, as Childs et al. [24] did, to avoid such risks.

3.17.2. The Supply Chain

Li-ion batteries have accelerated the energy storage market, thanks to cost declines achieved through utilisation in multiple industries, like EVs and consumer electronics [50]. However, an ongoing topic of research and discussion is not only the lack of transparency in the value chain of the raw materials (e.g., nickel, manganese, and cobalt) used in the technology but whether there are enough raw materials to satisfy the needs for all applications [4,34,51]. These concerns have initiated new energy storage developments and spurred further interest in the field [52].
Still, when looking at the reviewed literature on LDES, there is little conversation or risk reservation for sustainability or supply chain considerations, even if several of the sources include Li-ion batteries in the modelling. Many LDES technologies rely on existing supply chains using earth-abundant materials available in substantial quantities globally (in the core technology and the balance of plant systems). However, this is not the case for all LDES equipment, as some use rare-earth magnetic materials and other elements that face supply challenges in certain regions, such as vanadium, which has limited distribution in Europe. Consequently, there is potential for scarcity [4]. Therefore, this topic is raised as a variable that may influence LDES deployment negatively, even though it is not a variable causing divergence in results for the reviewed literature.
Controversially, the supply chain is also an important factor since a shortage of supply or uncertainties with Li-ion batteries and other balancing technologies could positively impact the deployment of LDES. An example is the uncertainty with natural gas used in conventional power plants since there is a lot of price fluctuation, and fuel availability can also be unreliable. Lack of predictability combined with a wide range of projected fuel prices (USD 8.50/MMBtu to USD 17.50/MMBtu in 2040) is bad for building a strong business case, which will ultimately affect the generation mix, dispatch order, and requirements on flexibility [3].

3.17.3. Reliability and Availability Constraints

Ziegler et al. [34] found that the levelized cost of shaped electricity (LCOSE) for LDES decreases by 25% when 95% availability can be assumed instead of 100%. Periods of unmet demand could then be met with DSM or supplemental generation like gas turbines or engines, which could be mitigation to avoid 100% availability requirements, thus improving the feasibility of the technologies [34].
Shaner et al. [29] also highlight that reliability requirements in the electricity system will impact the extent of solar and wind power contributions in the generation mix. For example, the North American Electricity Reliability Corporation (NERC) reliability standard specifies a loss of load expectation of 0.1 days per year, corresponding to 99.97% reliability [29]. The stricter this requirement is, the more critical flexibility and proper backup generation becomes, creating additional opportunities for LDES. This can be understood from the findings of Shaner et al. [29] since meeting 99.97% of total annual electricity demand in their model based on solar and wind power combined with 12 h of energy storage requires up to 2.2x generation, but when the energy storage capacity is increased to 32 days, the generation need is reduced to 1.1x the same generation mix. However, this reliability aspect was not raised as another variable in Table 3, as the standard in the literature is to assume the full demand is met (100% reliability).

3.17.4. Political and International Relations

Although not considered in the reviewed papers, politics and world order heavily influence the power sector. This has been evident in the Russia and Ukraine conflict, where restrained natural gas supply created turmoil in European Union (EU) markets [53]. This, in turn, has accelerated interest in hydrogen and self-sufficiency [54], directly linking LDES to the discussion, which could speed up the deployment of the technologies.

3.17.5. Safety Aspects

Managing safety aspects for specific LDES technologies will impact competitiveness, as such concerns will continue shaping regulators’ assessments of various technologies [13]. While this element’s effect on implementing LDES in practical settings may not be apparent in the analysed studies, it is still anticipated to have some influence.

3.17.6. Renewable Phrasing

Not all renewable energy is variable and dependent on weather. For instance, biomass is a sustainable and dispatchable resource that can complement the variability of solar and wind, as in Child et al. [2] and Gaete-Morales et al. [20]. Thus, distinguishing renewable energy (RE) from variable renewable energy (VRE) is essential since it can be observed from the literature that misinterpretations have been made. For example, Hunter et al. [32] use Zhang et al. [25] as a reference case for their modelling, consisting of 85% RE, while later stating that LDES will be needed at 80% VRE penetration. However, Zhang et al. [25] have 85% RE but only 61% VRE while utilising LDES technologies. Therefore, VRE and RE as expressions appear to have been misused here, as both models are based on the same foundation.
Consistency throughout one’s text in how renewable energy is addressed is critical to avoid confusion. A good example of such consistency is Blanco and Faaij [10]. Conversely, Bettoli et al. [4], Schill [11], and Bogdanov and Breyer [33] are examples of sources where the terms RE and VRE are used with varying meanings since the discussion occasionally involves variability, but the term renewable energy is used. Thus, the reader might misunderstand if only solar and wind power or other renewable sources are considered. In the current literature, LDES feasibility is communicated in association with RE and VRE penetration levels, which can cause misunderstandings in comparisons between studies. In Child et al. [2], for instance, it is visible that LDES will become needed at RE penetrations of 73% and above, while Denholm et al. [23] argue the same but at 50% VRE utilisation. In Child et al. [2], however, a 73% RE penetration also entails about 50% VRE usage. The same applies to Du et al. [1], where 70% RE penetration means 54.2% VRE and to Schill and Zerrahn [28], where 80% RE implies 70% VRE. In an LDES context, the VRE penetration level should be used (instead of RE) as the variability of these sources is of interest. All percentage values noted in Table 3 have been converted (where needed) to VRE for alignment purposes.

3.18. Overview

The previously examined collection of in total 21 variables can be categorised into three segments based on their influence on the researched phenomenon: modelling conditions, geographical circumstances, and system design, as illustrated in Figure 4.
Modelling conditions encapsulate variables related to grid simulation with connected inputs and assumptions, while geographical circumstances capture factors inherited or mainly affected by the chosen location. System design, in turn, involves elements that create the study’s framework.
To build on this review, additional analysis could be conducted to further specify the relationship between VRE penetration and LDES investments once the market is created since LDES and VRE capacity additions are unlikely to be linearly linked. This is a complex subject, so examining the impact of each of the 21 variables would be beneficial since some will have more influence than others. Much weight in the current literature is also placed on modelling various electricity systems, clarifying the role of LDES, and identifying needed policy and market mechanisms, but little attention is given to gathering input and opinions from actual investors or companies in the power markets. This could help recognise how research and academia could better support decarbonising the power sector and spur LDES implementation.
Another interesting reflection is that the LDES conversation tends to be initiated from the idea of a 100% renewable energy system, while it could be debated if this level of RE development is genuinely required or realistic. Shaner et al. [29] found that advancing from 80% of demand met by solar and wind power generation to 100% involves a substantial increase in power capacity and energy storage, while the added value is limited. Guerra et al. [5] and Ziegler et al. [34] also demonstrate a considerable cost increase as a system approaches a 100% renewable energy mix. Therefore, interconnections and other technologies, e.g., more conventional power generation combined with CCS or new nuclear power developments within small modular reactors (SMR), could be crucial in optimising the full system from a cost and emissions perspective. The strategy for feasibly achieving set emission targets while securing high enough energy reliability will ultimately influence the market size and potential for LDES.
The limitations of the research presented herein are intertwined with the chosen methodology, which involved conducting a literature review. During the process of gathering studies considered relevant to the research, it is important to acknowledge that complete coverage of the applicable literature is never guaranteed, thus resulting in potential gaps in the review. Furthermore, the review has time and language constraints, focusing on studies published in English from 2015 onward. Moreover, the analysis excludes studies behind certain paywalls and unpublished research at the time of concluding the review, thereby potentially neglecting additional input and alternative perspectives. Another factor contributing to a possible limitation is the inevitable varying quality of included studies, which pose challenges when comparing and synthesising results. However, all of these limitations were anticipated and therefore addressed through the research strategy of meticulously selecting a diverse array of high-quality databases adapted to enhance the inclusion of valuable information while reducing the likelihood of missing important data.

4. Conclusions

The level of VRE penetration in an energy system is a critical parameter when evaluating the commercial feasibility of LDES. This review set out to enhance the understanding of how this correlation is distinguishable in the existing literature, to enable further exploration of the nuanced connection between VRE integration and the development of LDES markets. A central finding of the research is the substantial diversity in the prevailing literature’s approaches to grid modelling in the context of LDES utilisation, leading to broad variations in outcomes related to the linkage between VRE deployment and the adoption of LDES technologies. This finding underscores the challenge of establishing a single, universally applicable level of VRE penetration that could accurately serve as an indicator of LDES market creation, even though such VRE penetration levels have been proposed in past studies.
To clarify the reasons for the broad variations in outcomes, a comprehensive framework of 16 variables was created. These variables, identified through an extensive review of the literature, are known to exert influence on the relationship between VRE penetration and LDES technology adoption. Furthermore, this study has discovered an additional five variables, which are presented in the supplementary findings section. As such, this research contributes to a deeper understanding of the complex link between VRE integration and the evolution of LDES markets, shedding light on critical aspects that previous reviews have not thoroughly examined. By explaining the determinants affecting the adoption of LDES within diverse contexts, this in-depth examination enriches the comprehension of LDES market creation. Additionally, it identifies less explored areas within the field that warrant further investigation. Another contribution of this review is the discovery of imprecise terminology associated with renewable energy, as the value proposition of LDES relates to VRE installations and not the broader spectrum of RE. This vagueness can cause unnecessary confusion. It has also been realised that most grid modelling tends to be performed for one-year periods but as highlighted, this might not properly indicate the value delivered by LDES technologies.
As a result, this study offers valuable insights for both scholars and practitioners who aim to navigate the evolving landscape of LDES utilisation. Policymakers should prioritise clarity in renewable energy terminology to avoid confusion and facilitate effective communication and policy implementation for LDES technologies. Instead of focusing on a specific VRE penetration level, they should recognise the variability across systems and emphasise accurate modelling to determine optimal timing for LDES deployment, incorporating the variables identified in this study. Energy planners need to consider a broader range of variables when assessing the feasibility and installation of LDES technologies, adopting a comprehensive approach to evaluate benefits and challenges under different VRE penetration scenarios. Extending modelling horizons beyond the typical one-year period is also crucial to capture the long-term value and performance of LDES. For researchers, this study underscores the importance of more standardised methodologies and consistent terminology in evaluating LDES and VRE integration. The identification of 16 key variables, along with five additional factors, provides a robust framework for future research, encouraging deeper exploration of these variables and their interactions to develop more accurate and reliable models.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/en17153779/s1, Supplementary File S1: outline of the reviewed literature.

Author Contributions

Conceptualisation, A.S.; methodology, A.S.; formal analysis, A.S.; data curation, A.S.; writing—original draft preparation, A.S.; writing—review and editing, M.H. and M.B.-S.; visualisation, A.S.; supervision, M.H. and M.B.-S.; project administration, A.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research was performed as part of employment by Wärtsilä Finland Oy and received separate funding from Högskolestiftelsen i Österbotten, grant number 286001114K1, Aktiastiftelsen i Vasa, grant number 28600129K1, and the APC was funded by Gösta Branders research fund, Åbo Akademi Research Foundation.

Data Availability Statement

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

Acknowledgments

The researchers acknowledge their gratitude towards Wärtsilä Finland Oy for guidance in the research.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Du, E.; Jiang, H.; Xiao, J.; Hou, J.; Zhang, N.; Kang, C. Preliminary analysis of long-term storage requirement in enabling high renewable energy penetration: A case of East Asia. IET Renew. Power Gener. 2021, 15, 1255–1269. [Google Scholar] [CrossRef]
  2. Child, M.; Kemfert, C.; Bogdanov, D.; Breyer, C. Flexible electricity generation, grid exchange and storage for the transition to a 100% renewable energy system in Europe. Renew. Energy 2019, 139, 80–101. [Google Scholar] [CrossRef]
  3. Kaushik, E.; Prakash, V.; Mahela, O.P.; Khan, B.; El-Shahat, A.; Abdelaziz, A.Y. Comprehensive Overview of Power System Flexibility during the Scenario of High Penetration of Renewable Energy in Utility Grid. Energies 2022, 15, 516. [Google Scholar] [CrossRef]
  4. Bettoli, A.; Linder, M.; Nauclér, T.; Noffsinger, J.; Sengupta, S.; Tai, H.; van Gendt, G. Net-Zero Power: Long-Duration Energy Storage for a Renewable Grid; LDES Council, McKinsey & Company: Brussels, Belgium, 2021. [Google Scholar]
  5. Guerra, O.J.; Eichman, J.; Denholm, P. Optimal energy storage portfolio for high and ultrahigh carbon-free and renewable power systems. Energy Environ. Sci. 2021, 14, 5132–5146. [Google Scholar] [CrossRef]
  6. de Sisternes, F.J.; Jenkins, J.D.; Botterud, A. The value of energy storage in decarbonizing the electricity sector. Appl. Energy 2016, 175, 368–379. [Google Scholar] [CrossRef]
  7. Ho, C.K.; McNamara, J.W. The Value of Long-Duration Energy Storage: Policy and Perception; Sandia National Laboratories: Albuquerque, NM, USA, 2021. [Google Scholar]
  8. Denholm, P.; Cole, W.; Frazier, A.W.; Podkaminer, K.; Blair, N. The Challenge of Defining Long-Duration Energy Storage; National Renewable Energy Laboratory: Golden, CO, USA, 2021. [Google Scholar]
  9. Zingales, A. Smart Storage the Key factor of Energy Transition. IOP Conf. Ser. Earth Environ. Sci. 2022, 1073, 012014. [Google Scholar] [CrossRef]
  10. Blanco, H.; Faaij, A. A review at the role of storage in energy systems with a focus on Power to Gas and long-term storage. Renew. Sustain. Energy Rev. 2018, 81, 1049–1086. [Google Scholar] [CrossRef]
  11. Schill, W.-P. Electricity Storage and the Renewable Energy Transition. Joule 2020, 4, 2059–2064. [Google Scholar] [CrossRef]
  12. Albertus, P.; Manser, J.S.; Litzelman, S. Long-Duration Electricity Storage Applications, Economics, and Technologies. Joule 2020, 4, 21–32. [Google Scholar] [CrossRef]
  13. Ratz, H.; Robichaud, R.; Bird, L.; Hutchinson, N. The Role of Long-Duration Energy Storage in Deep Decarbonization: Policy Considerations; World Resources Institute: Washington, DC, USA, 2020. [Google Scholar] [CrossRef]
  14. Bertsch, J.; Growitsch, C.; Lorenczik, S.; Nagl, S. Flexibility in Europe’s power sector—An additional requirement or an automatic complement? Energy Econ. 2016, 53, 118–131. [Google Scholar] [CrossRef]
  15. Dowling, J.A.; Rinaldi, K.Z.; Ruggles, T.H.; Davis, S.J.; Yuan, M.; Tong, F.; Lewis, N.S.; Caldeira, K. Role of Long-Duration Energy Storage in Variable Renewable Electricity Systems. Joule 2020, 4, 1907–1928. [Google Scholar] [CrossRef]
  16. Jacobson, M.Z.; Delucchi, M.A.; Cameron, M.A.; Frew, B.A. Low-cost solution to the grid reliability problem with 100% penetration of intermittent wind, water, and solar for all purposes. Proc. Natl. Acad. Sci. USA 2015, 112, 15060–15065. [Google Scholar] [CrossRef]
  17. Helman, U.; Kaun, B.; Stekli, J. Development of Long-Duration Energy Storage Projects in Electric Power Systems in the United States: A Survey of Factors Which Are Shaping the Market. Front. Energy Res. 2020, 8, 539752. [Google Scholar] [CrossRef]
  18. Leonard, M.D.; Michaelides, E.E.; Michaelides, D.N. Energy storage needs for the substitution of fossil fuel power plants with renewables. Renew. Energy 2020, 145, 951–962. [Google Scholar] [CrossRef]
  19. Cole, W.; Denholm, P.; Carag, V.; Frazier, W. The peaking potential of long-duration energy storage in the United States power system. J. Energy Storage 2023, 62, 106932. [Google Scholar] [CrossRef]
  20. Gaete-Morales, C.; Gallego-Schmid, A.; Stamford, L.; Azapagic, A. A novel framework for development and optimisation of future electricity scenarios with high penetration of renewables and storage. Appl. Energy 2019, 250, 1657–1672. [Google Scholar] [CrossRef]
  21. Bryan, J.; Meek, A.; Dana, S.; Islam Sakir, M.S.; Wang, H. Modeling and design optimization of carbon-free hybrid energy systems with thermal and hydrogen storage. Int. J. Hydrogen Energy 2023, 48, 39097–39111. [Google Scholar] [CrossRef]
  22. Guerra, O.J.; Zhang, J.; Eichman, J.; Denholm, P.; Kurtz, J.; Hodge, B.-M. The value of seasonal energy storage technologies for the integration of wind and solar power. Energy Environ. Sci. 2020, 13, 1909–1922. [Google Scholar] [CrossRef]
  23. Denholm, P.; Cole, W.; Frazier, A.W.; Podkaminer, K.; Blair, N. The Four Phases of Storage Deployment: A Framework for the Expanding Role of Storage in the U.S. Power System; National Renewable Energy Laboratory: Golden, CO, USA, 2021. [Google Scholar]
  24. Childs, E.; Roumpani, M.; Dueñas, S.; Sanchez, P.; Gorman, J.; Davidson, M.; Backer, L. Long Duration Energy Storage for California’s Clean, Reliable Grid; Strategen: Sacramento, CA, USA, 2020. [Google Scholar]
  25. Zhang, J.; Guerra, O.J.; Eichman, J.; Pellow, M.A. Benefit Analysis of Long-Duration Energy Storage in Power Systems with High Renewable Energy Shares. Front. Energy Res. 2020, 8, 527910. [Google Scholar] [CrossRef]
  26. Ganzer, C.; Pratama, Y.W.; Mac Dowell, N. The role and value of inter-seasonal grid-scale energy storage in net zero electricity systems. Int. J. Greenh. Gas Control 2022, 120, 103740. [Google Scholar] [CrossRef]
  27. Nikolakakis, T.; Broekman, E.B.; Chattopadhyay, D.M.; Merino, A.M. Analysis of Long-Term Variable Renewable Energy Heavy Capacity Plans Including Electric Vehicle and Hydrogen Scenarios: Methodology and Illustrative Case Study for Turkey. IEEE Access 2023, 11, 27189–27216. [Google Scholar] [CrossRef]
  28. Schill, W.-P.; Zerrahn, A. Long-run power storage requirements for high shares of renewables: Results and sensitivities. Renew. Sustain. Energy Rev. 2018, 83, 156–171. [Google Scholar] [CrossRef]
  29. Shaner, M.R.; Davis, S.J.; Lewis, N.S.; Caldeira, K. Geophysical constraints on the reliability of solar and wind power in the United States. Energy Environ. Sci. 2018, 11, 914–925. [Google Scholar] [CrossRef]
  30. Weitemeyer, S.; Kleinhans, D.; Vogt, T.; Agert, C. Integration of Renewable Energy Sources in future power systems: The role of storage. Renew. Energy 2015, 75, 14–20. [Google Scholar] [CrossRef]
  31. Belderbos, A.; Delarue, E.; D’haeseleer, W. Possible role of power-to-gas in future energy systems. In Proceedings of the 2015 12th International Conference on the European Energy Market (EEM), Lisbon, Portugal, 19–22 May 2015; pp. 1–5. [Google Scholar] [CrossRef]
  32. Hunter, C.A.; Penev, M.M.; Reznicek, E.P.; Eichman, J.; Rustagi, N.; Baldwin, S.F. Techno-economic analysis of long-duration energy storage and flexible power generation technologies to support high-variable renewable energy grids. Joule 2021, 5, 2077–2101. [Google Scholar] [CrossRef]
  33. Bogdanov, D.; Breyer, C. The Role of Solar Energy towards 100% Renewable Power Supply for Israel: Integrating Solar PV, Wind Energy, CSP and Storages. In Proceedings of the 19th Sede Boqer Symposium on Solar Electricity Production, Beersheba, Israel, 23–25 February 2015. [Google Scholar]
  34. Ziegler, M.S.; Mueller, J.M.; Pereira, G.D.; Song, J.; Ferrara, M.; Chiang, Y.-M.; Trancik, J.E. Storage Requirements and Costs of Shaping Renewable Energy Toward Grid Decarbonization. Joule 2019, 3, 2134–2153. [Google Scholar] [CrossRef]
  35. Mannina, G.; Ekama, G.; Caniani, D.; Cosenza, A.; Esposito, G.; Gori, R.; Garrido-Baserba, M.; Rosso, D.; Olsson, G. Greenhouse gases from wastewater treatment—A review of modelling tools. Sci. Total Environ. 2016, 551–552, 254–270. [Google Scholar] [CrossRef] [PubMed]
  36. Lund, P.D.; Lindgren, J.; Mikkola, J.; Salpakari, J. Review of energy system flexibility measures to enable high levels of variable renewable electricity. Renew. Sustain. Energy Rev. 2015, 45, 785–807. [Google Scholar] [CrossRef]
  37. Ministry of Foreign Affairs of Israel. Israel-Size and Dimension. Isr Missions World n.d. Available online: https://embassies.gov.il/MFA/Pages/mfa.aspx (accessed on 5 February 2023).
  38. Statistics Times. List of Countries by Area. Stat Times. 2020. Available online: https://statisticstimes.com/geography/countries-by-area.php (accessed on 5 February 2023).
  39. Armstrong, R.C.; Barbar, M.; Brown, P.; Brushett, F.; Chiang, Y.-M.; Deutch, J.; Gencer, E.; Gruenspecht, H.; Joskow, P.; Junge, C.; et al. The Future of Energy Storage; Massachusetts Institute of Technology: Cambridge, MA, USA, 2022. [Google Scholar]
  40. McNamara, J.W.; DeAngelis, V.; Byrne, R.H.; Benson, A.; Chalamala, B.R.; Masiello, R. Long-duration energy storage in a decarbonized future: Policy gaps, needs, and opportunities. MRS Energy Sustain. 2022, 9, 142–170. [Google Scholar] [CrossRef]
  41. Sánchez-Pérez, P.A.; Staadecker, M.; Szinai, J.; Kurtz, S.; Hidalgo-Gonzalez, P. Effect of modeled time horizon on quantifying the need for long-duration storage. Appl. Energy 2022, 317, 119022. [Google Scholar] [CrossRef]
  42. Sargent & Lundy. Combined-Cycle Plant Life Assessment; Sargent & Lundy: Chicago, IL, USA, 2017. [Google Scholar]
  43. Sepulveda, N.A.; Jenkins, J.D.; de Sisternes, F.J.; Lester, R.K. The Role of Firm Low-Carbon Electricity Resources in Deep Decarbonization of Power Generation. Joule 2018, 2, 2403–2420. [Google Scholar] [CrossRef]
  44. Poli, N.; Trovò, A.; Fischer, P.; Noack, J.; Guarnieri, M. Electrochemical rebalancing process for vanadium flow batteries: Sizing and economic assessment. J. Energy Storage 2023, 58, 106404. [Google Scholar] [CrossRef]
  45. Denholm, P.; Mai, T. Timescales of Energy Storage Needed for Reducing Renewable Energy Curtailment. Renew. Energy 2017, 130, 388–399. [Google Scholar] [CrossRef]
  46. Pudjianto, D.; Strbac, G. Whole system value of long-duration electricity storage in systems with high penetration of renewables. IEnergy 2022, 1, 114–123. [Google Scholar] [CrossRef]
  47. Schaber, K.; Steinke, F.; Mühlich, P.; Hamacher, T. Parametric study of variable renewable energy integration in Europe: Advantages and costs of transmission grid extensions. Energy Policy 2012, 42, 498–508. [Google Scholar] [CrossRef]
  48. Steinke, F.; Wolfrum, P.; Hoffmann, C. Grid vs. storage in a 100% renewable Europe. Renew. Energy 2013, 50, 826–832. [Google Scholar] [CrossRef]
  49. Strbac, G.; Aunedi, M.; Pudjianto, D.; Djapic, P.; Teng, F.; Sturt, A.; Jackravut, D.; Sansom, R.; Yufit, V.; Brandon, N. Strategic Assessment of the Role and Value of Energy Storage Systems in the UK Low Carbon Energy Future; Imperial College London: London, UK, 2012. [Google Scholar]
  50. Edington, A.N.C. The Role of Long Duration Energy Storage in Decarbonizing Power Systems; Massachusetts Institute of Technology: Cambridge, MA, USA, 2019. [Google Scholar]
  51. Azevedo, M.A.; Campagnol, N.; Hagenbruch, T.; Hoffman, K.; Lala, A.; Ramsbottom, O. Lithium and Cobalt—A Tale of Two Commodities; McKinsey & Company: Chicago, IL, USA, 2018. [Google Scholar]
  52. Lee, A. The Trouble with Lithium; Bloomberg: New York, NY, USA, 2022. [Google Scholar]
  53. European Council. Impact of Russia’s Invasion of Ukraine on the Markets: EU Response; Eur Counc Counc Eur Union: Brussels, Belgium, 2023; Available online: https://www.consilium.europa.eu/en/policies/eu-response-ukraine-invasion/impact-of-russia-s-invasion-of-ukraine-on-the-markets-eu-response/ (accessed on 5 February 2023).
  54. International Energy Agency. Russia’s War on Ukraine; IEA: Paris, France, 2023; Available online: https://www.iea.org/topics/russias-war-on-ukraine (accessed on 5 February 2023).
Energies 17 03779 g001
Figure 1. Literature review process, see Supplementary Materials—File S1 for sources.
Figure 1. Literature review process, see Supplementary Materials—File S1 for sources.
Energies 17 03779 g001
Energies 17 03779 g002
Figure 2. Visualisation of how the reviewed literature indicates LDES market creation in relation to VRE penetration levels. Sources that conduct independent modelling are underscored.
Figure 2. Visualisation of how the reviewed literature indicates LDES market creation in relation to VRE penetration levels. Sources that conduct independent modelling are underscored.
Energies 17 03779 g002
Energies 17 03779 g003
Figure 3. Timescales of different energy storage technologies ([1]; modified).
Figure 3. Timescales of different energy storage technologies ([1]; modified).
Energies 17 03779 g003
Energies 17 03779 g004
Figure 4. Recognised variables categorised into applicable segments. White boxes represent additional variables, relevant to the subject matter, but found to have no impact on the analysed results or were unaddressed in the reviewed literature.
Figure 4. Recognised variables categorised into applicable segments. White boxes represent additional variables, relevant to the subject matter, but found to have no impact on the analysed results or were unaddressed in the reviewed literature.
Energies 17 03779 g004
Table 1. Utilised databases and noted search results.
Table 1. Utilised databases and noted search results.
DatabaseSearch Result
Alma (Åbo Akademi University)370
Google Scholar193
Academic Search Complete (EBSCO)149
Energy and Power Source (EBSCO)31
IEEE/IEE Electronic Library—IEL31
Scopus145
Web of Science138
The electronic search string was identical for all databases: (abstract: “long duration energy storage” OR “long-duration energy storage” OR “long term energy storage” OR “long-term energy storage” OR “seasonal energy storage” OR “diurnal energy storage” OR “weekly energy storage” OR “monthly energy storage” OR “inter-seasonal energy storage” OR “multi-day energy storage” OR “multi day energy storage”).
Table 2. Relevance of the 16 identified variables.
Table 2. Relevance of the 16 identified variables.
VariableRelevance/Inclusion Criteria
Estimated Costs and EfficienciesKey drivers for LDES deployment in electricity grid system modelling.
LDES Value and Business Synergies
Applied Emission Requirements
Assumed VRE Penetration Level
Assumed InterconnectionsFactors influencing the required flexibility in an electricity grid system, directly impacting LDES deployment.
Demand Side Management
Consideration of Past Installations
Weather Conditions
VRE Mix
Curtailment Possibilities
Modelling Tool and PurposeVariables that limit or frame the outcomes of a study, contributing to varying results in the literature.
Model Horizon and Approach
Geographical Market
Assumed Electricity Demand
Available Balancing Technologies (LDES Type)
Storage Duration
LDES: Long-duration energy storage; VRE: Variable renewable energy.
Table 3. Summary of the reviewed literature’s approach to the 16 identified variables.
Table 3. Summary of the reviewed literature’s approach to the 16 identified variables.
SourceVRE [%]Modelling
Tool		MarketPast
Installations		Emission
Requirements		Cost
and
Efficiency		Multiple
Balancing Technologies
(LDES Type)		Inter
Connections		LDES Value
and
Business		Model Horizon [Years]VRE Mix [%]Weather ConditionsIncreased Electricity
Demand		DSMStorage Duration
[hours]		Curtailment
[17]20 No modelUSA-----------≥8–12-
[18]25 NAERCOT (USA)X-----177W/23SX--Seasonal -
[19]37ReEDSUSAX--XX-222W
/78S		XX-4–168-
[20]40 FuturES,
power-GAMA		ChileXXXX--1 50W/50SXX-17X
[21]47Dymola,
Modelica,
SciPy Python		ERCOT (USA)--X---142W/58SX--3 days-
[22]49 ReEDS, PLEXOS,
RODeO		Western U.S. Power SystemX-XXXX1 55W/45SXX-≥24 X
[2]50 LUT Energy System Transition modelEUX-XX--5 65W/35SXXX24–100NA
[14]50DIMEN-SIONItalyXXXXX-1 58W/42SXXXNAX
[23]50 No modelUSA-----------4–12-
[24]50 GridPath,
RESOLVE		CAISO (USA)XXXXXX1 15W/85SXXX10X
[1]54 Generation
Capacity
Planning Model		East AsiaX-XXNA-1 45W/55SXX-720X
[11]60 No modelGermany-----------11.3-
[4]60 McKinsey Power Model (MPM), McKinsey
Battery Cost Model, McKinsey
Energy
Insights		GlobalXXXXXXNANAXXNA8–150X
[5]60 SDOMCAISO (USA)X-XX--1 43W/57SXX-4.9–1404X
[25]61 PLEXOS, ReEDS Western U.S. Power SystemX-XX-X1 54W/46SXX-10X
[26]64ESO-XUKXXX-X-161W/39SXX-8400X
[27]70EPMTurkeyXXXX--260W/40SXX-SeasonalX
[28]70 DIETERGermany-XXX--1 50W/50SX-XNAX
[6]70 IMRES ERCOT (USA)-XX---1 71W/29SXX-10X
[7]70 No modelUSA/
Global		----------->4 -
[12]70 No modelUSA-----------10–100-
[9]70 No modelGlobal-----------Days or weeks-
[14]72 *DIMENSIONEUXXXXX-1 69W/31SXXXNA X
[5]73 SDOMMISO (USA)X-XX--1 84W/16SXX-4.9–1404X
[29]80 Simple
Transparent Model		USA----X-1 25W/75SX--12X
[30]80 NAGermany------8 60W/40SX--168X
[3]80 No modelGlobal-----------NA-
[31]80 Linear
Program (LP)
Investment Model		Belgium-XX---1 100WX--SeasonalX
[13]80 No modelGlobal-----------Seasonal-
[8]80 No modelUSA/
Global		-----------10-
[32]80 StoreFASTWestern U.S. Power System--XX--1 NAXX-12–168-
[33]85 LUT Linear
Optimisation Model		IsraelXNAX-NA-1 29W/71SXXXSeasonalX
[34]95 System
Advisor Model, WRF Model		Arizona, Massa-chusetts, Iowa, Texas (USA)--X---20 NAX--6–180X
[16]100 LOAD-MATCHUSA--XXX-6 52W/48SXXXSeveral weeks or months-
[15]100 Macro Scale
Energy Model		USA-XX-X -6 NAX-X≥10 X
VRE [%] *: Good example of where flexible natural gas power plants are supporting a certain VRE %, following the same typical operating profile as LDES, enabled through low CO2 emission requirements; No model: The source is not modelling or optimising a specific electricity grid; X: Included; -: Not included; NA: Not answered; Storage Duration: How the source defines LDES, uses the expression or how LDES technologies are being utilised in the model. The requirement for the source to be included was energy storage of 10 h or longer; DSM: Demand side management; and W/S: Wind/Solar.
Table 4. Modelling tools and purposes for the reviewed literature.
Table 4. Modelling tools and purposes for the reviewed literature.
SourceModelling ToolPurpose
[1]Generation Capacity Planning ModelTo quantitatively analyse the role of long-term seasonal storage in enabling high VRE penetration
[2]LUT Energy System Transition modelTo optimise energy system elements to minimise total annualised system costs and the cost of end-user electricity consumption
[4]McKinsey Power Model (MPM), McKinsey Battery Cost Model, McKinsey Energy InsightsTo establish the role of LDES solutions in electrical power systems
[5]SDOMTo assess an optimal storage portfolio based on variable renewable power deployment
[6]IMRESTo explore cost optimisation and the potential value of energy storage in deep decarbonisation of the electricity sector
[14]DIMENSIONTo analyse if there is a need for additional incentive mechanisms for flexibility in electricity markets with high shares of renewables
[25]PLEXOS, ReEDSTo explore system-level services and associated benefits of long-duration energy storage
[24]GridPath, RESOLVETo improve on previous modelling approaches to better reflect the capabilities and value of long-duration energy storage resources
[18]NATo calculate the implications of substituting fossil fuel power plants with renewable energy sources
[15]Macro Scale Energy ModelTo assess the potential of long-duration energy storage technologies to enable reliable and cost-effective VRE-dominated electricity systems
[34]System Advisor Model, WRF ModelTo investigate if energy storage can cost-competitively shape intermittent resources into desired output profiles and compare diverse storage technologies
[30]NATo analyse the influence of storage size and efficiency on the pathway towards a 100% renewable energy scenario
[29]Simple Transparent ModelTo quantify the coverability of solar and wind resources as a function of time and location over multi-decadal time scales and up to continental length scales
[28]DIETERTo analyse the role of power storage in systems with high shares of variable renewable energy sources
[20]FuturES and powerGAMATo present a new framework for developing future electricity scenarios with a high penetration of renewables
[31]Linear Program (LP) Investment ModelTo compare the possible opportunities of power-to-gas as a long-term storage option, to the opportunities of short-term storage technologies
[22]ReEDS, PLEXOS, RODeOTo propose a model-based approach for comprehensive techno-economic assessments of grid-integrated seasonal storage
[32]StoreFASTTo provide a detailed techno-economic evaluation and an uncertainty analysis of applicable technologies, as well as identify challenges and opportunities to support electric grid planning
[33]LUT Linear Optimisation ModelTo compute an optimal mix of technologies for different shares of renewable energy and define the cost of electricity for every configuration of the energy system
[16]LOADMATCHTo address the high cost of avoiding load loss caused by variability and uncertainty of wind, water, and solar power
[19]ReEDSTo evaluate the peaking potential of storage, meaning its ability to substitute for the traditional capacity resources used to ensure resource adequacy
[26]ESO-XTo explore the potential for inter-seasonal energy storage in the context of a net zero energy system
[27]EPMTo minimise total system costs while investigating how EV load and green hydrogen can be integrated into long-term electricity models
[21]Dymola, Modelica, SciPy PythonTo examine the feasibility of replacing all coal and natural gas electricity generation in the state of Texas with solar, wind, and nuclear power, combined with thermal energy and hydrogen storages
NA: Not answered.
Table 5. Comparable electricity demand estimates in the reviewed literature.
Table 5. Comparable electricity demand estimates in the reviewed literature.
RegionSourceDemand Estimate [TWh]VRE Mix [%]
Western U.S. Power System[22]1122 (2050)55W/45S
Western U.S. Power System[25]1128 (2050)54W/46S
Western U.S. Power System[32]1124 (2050)NA
CAISO (USA)[24]266 (2030)15W/85S
CAISO (USA)[5]269 (2050)43W/57S
EU[2]5116 (2050)65W/35S
EU[14]4171 (2050)58W/42S
NA: Not answered.
Table 6. Applied emission requirements in the reviewed literature.
Table 6. Applied emission requirements in the reviewed literature.
SourcePeriodEmission Requirement
[20]2015–20295 $/t CO2
2030–205010 $/t CO2
[26]2020–205018–236 £/t CO2
[14]2020–2050 (scenario A)22.6–50.0 €/t CO2
2020–2050 (scenario B)35.1–100.0 €/t CO2
[4]2030 (base case)60 $/tCO2 + 8% compound annual growth rate 2030–2040
2030 (medium case)75 $/tCO2 + 8% compound annual growth rate 2030–2040
2030 (high case)100 $/tCO2 + 8% compound annual growth rate 2030–2040
[28]2050100 €/t CO2
[24]203030 MMT CO2
20450 MMT CO2
[27]205035 MMT CO2
[6]203550–200 tCO2/GWh
[31]2013Carbon neutral system
[15]2050Carbon neutral system
MMT: Million metric tons.
Table 7. Summary of considered LDES technologies in the reviewed literature, with related assumptions. Special remarks highlighted with brackets.
Table 7. Summary of considered LDES technologies in the reviewed literature, with related assumptions. Special remarks highlighted with brackets.
SourceLDES TechnologiesPower Capacity CostEnergy Capacity CostCost BaseEfficiency
[32]Li-ion246 $/kW320 $/kWhYear 201886%
ETH1068 $/kW0.23 $/kWh52% [Discharge efficiency]
P2G (H2)1503 $/kW [Electrolysis]1–14.8 $/kWh37%
39 $/kW [Compressor]
130 $/kW [Rectifier]
VRB1384 $/kW196 $/kWh75%
TES1670 $/kWNA52%
(NG + CCS)2411 $/kWNA48% [Discharge efficiency]
PHES1612 $/kW83 $/kWh81%
CAES507–759 $/kW18–27 $/kWh55–65%
[22]PHES573–2807 $/kW17.3–97.4 $/kWhYear 205080%
CAES415–947 $/kW8.9–81.6 $/kWh60%
P2G (H2)650–1950 $/kW0.5–1.5 $/kWh40%
[16]CSPNA15.3 $/kWhYear 205099%
PHESNA14 $/kWh80%
[20]CSP9000 $/kW46 $/MWhYear 2050NA
Hydro reservoirs2200 $/kW20 $/MWh
[Base storage value]
[2]CAESNA
248 €/kW [Electrolysis]
154 €t/CO2 [CO2 direct air capture]
190 €/kW [Methanation]		26.3 €/kWhYear 205084%
P2G (SNG)0.05 €/kWhNA
[14]CSP2805 €/kWNAYear 2050NA
(NG + CCS)1030–1314 €/kWNA33–52%
[28]PHES1100 €/kW10 €/kWhYear 205080%
P2G (H2)1000 €/kW0.2 €/kWh46%
[27]PHES0.92 $million/MW15,900 $/MW/year [FOM]
0 $/MWh [VOM]
8000 $/MW [FOM]
0 $/MWh [VOM]
58,000 $/MW/year [FOM]
2 $/MWh [VOM]		Year 203578%
P2G (H2)0.2 $million/MW [Electrolysis]Year 205080% [Electrolysis]
(NG + CCS)2.5 $million/MWYear 20356.8 mmBTU/MWh [Heat rate]
[21]P2G (H2)NA50 $/MWh [Electrolysis]NA83% [Electrolysis]
20 $/MWh [H2 storage]
[LCOE values]
[26]P2G (SNG)2400 £/kWNAYear 205029% [P2G2P]
[32]P2G (SNG)1500 €/kW0 €/kWhYear 201360%
[33]P2G (SNG)380 €/kW [Electrolysis]0.05 €/kWhYear 203084% [Electrolysis]
356 €/kW [CO2 from air]78% [CO2 from air]
234 €/kW [Methanation]77% [Methanation]
[15]P2G (H2)1058 $/kW0.16 $/kWhYear 205049%
[24]Generic model (several LDES options)10.1–12.4 $/MW0.39–0.64 $/MWhYear 204581%
6 $/MW0.25 $/MWh72%
7.5 $/MW
[Annualised all-inclusive costs]		0.125 $/MWh
[Annualised all-inclusive costs]		64%
[1]Generic model (several LDES options)600 $/kW63.2 $/MWhYear 205050%
800 $/kW62.3 $/MWh50%
1500 $/kW61.9 $/MWh50%
2400 $/kW62.6 $/MWh
[LCOE values]		50%
[4]Generic model (several LDES options)380–960 $/kW4–17 $/kWhYear 204055%
[5]Generic model (several LDES options)842.6 $/kW34.7 $/kWhYear 205063%
1063.3 $/kW53.1 $/kWh78.2%
1414.8 $/kW1.1 $/kWh44%
[25]Generic model (several LDES options)NANAYear 205040%
NANA60%
NANA70%
NANA80%
[19]Generic model (several LDES options)---35%
---60%
---85%
[6]Generic model (one LDES option)NA100–250 $/kWhYear 203580%
[29]Generic model (one LDES option)---NA
[30]Generic model (one LDES option)---80%
[34]Generic model (one LDES option)1000 $/kW20 $/kWhYear 202275%
CSP: Concentrated solar power; ETH: ethanol fuel utilization; LCOE: levelized cost of energy; NA: not answered; NG + CCS: natural gas utilisation with carbon capture and storage (not considered an LDES technology but very comparable use case); and P2G (H2): variable renewable energy converted and stored as hydrogen. Dispatch to generate electricity not included here (P2G2P); P2G (SNG): variable renewable energy converted and stored as synthetic natural gas. Dispatch to generate electricity not included here, unless otherwise noted (P2G2P); TES: thermal energy storage; VRB: vanadium redox flow battery; FOM: fixed operation and maintenance cost; VOM: variable operation and maintenance cost; and -: not considered.
Table 8. Grid services’ relationship to energy storage duration ([36]; modified).
Table 8. Grid services’ relationship to energy storage duration ([36]; modified).
Storage Duration [h]ServicesExamples of Technologies
≤0.1Power quality, regulationFlow batteries, flywheels, DSM
0.1–1Spinning reserve, contingency reserve, black startFlow batteries, PHES, DSM
1–72Load following, load
levelling/peak shaving/valley filling, transmission curtailment prevention, transmission loss
reduction, unit commitment		Flow batteries, CAES, PHES, DSM
>730Seasonal shiftingCAES, PHES
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Selänniemi, A.; Hellström, M.; Björklund-Sänkiaho, M. Long-Duration Energy Storage—A Literature Review on the Link between Variable Renewable Energy Penetration and Market Creation. Energies 2024, 17, 3779. https://doi.org/10.3390/en17153779

AMA Style

Selänniemi A, Hellström M, Björklund-Sänkiaho M. Long-Duration Energy Storage—A Literature Review on the Link between Variable Renewable Energy Penetration and Market Creation. Energies. 2024; 17(15):3779. https://doi.org/10.3390/en17153779

Chicago/Turabian Style

Selänniemi, Andreij, Magnus Hellström, and Margareta Björklund-Sänkiaho. 2024. "Long-Duration Energy Storage—A Literature Review on the Link between Variable Renewable Energy Penetration and Market Creation" Energies 17, no. 15: 3779. https://doi.org/10.3390/en17153779

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop