The role of electricity market design for energy storage in cost-efficient decarbonization

Abstract

•Storage operations play an increasingly important role as power systems decarbonize•Market participation choices of storage have asymmetrical effects on cost and emission•Market design for storage affects up to 53% of consumer cost and 16% of emission reduction•Future market designs must balance storage’s economic and emission impacts Energy storage is key to decarbonize power systems by allowing excess renewable energy to be stored and released back to the grid as needed. Ideally, storage should be charged from carbon-free and low-cost renewables and discharged to replace dirty and expensive fossil-fuel generation. However, in reality, energy storage participates in electricity markets with a profit-driven motive, its impact on reducing system costs or emissions is dependent on market design and storage's participation choices. In some cases, storage may increase system costs or emissions if the market design or incentives are not aligned with renewable and storage capacity present in the system. This study aims to evaluate how market designs can affect the contribution of energy storage to electricity economics and decarbonization, from early to deep decarbonization stages. The proposed open-source framework can be used by researchers and policymakers to assess emerging technologies and policy incentives. Energy storage is widely recognized by power system utilities and regulators as a crucial resource for achieving energy decarbonization. However, in deregulated power systems, investor-owned storage participates in electricity markets with a profit-driven motive. The alignment of such profit-driven operations with social welfare critically depends on market design and storage’s participation choices. This study employs an agent-based approach and investigates the impact of different market participation options on storage’s contribution to reducing electricity costs and carbon emissions. Our findings suggest that the existing electricity pool market design in North America may encourage early-stage storage adoptions but hinder progress toward deep decarbonization. We found that day-ahead markets are more effective in utilizing storage to reduce carbon emissions, while real-time markets are more effective in reducing costs. We compare different combinations of storage market participation choices and conclude trade-offs between consumer energy affordability and carbon emissions. Energy storage is widely recognized by power system utilities and regulators as a crucial resource for achieving energy decarbonization. However, in deregulated power systems, investor-owned storage participates in electricity markets with a profit-driven motive. The alignment of such profit-driven operations with social welfare critically depends on market design and storage’s participation choices. This study employs an agent-based approach and investigates the impact of different market participation options on storage’s contribution to reducing electricity costs and carbon emissions. Our findings suggest that the existing electricity pool market design in North America may encourage early-stage storage adoptions but hinder progress toward deep decarbonization. We found that day-ahead markets are more effective in utilizing storage to reduce carbon emissions, while real-time markets are more effective in reducing costs. We compare different combinations of storage market participation choices and conclude trade-offs between consumer energy affordability and carbon emissions. Grid-scale battery energy storage (“storage”) contributes to a cost-efficient decarbonization process provided that it charges from carbon-free and low-cost renewable sources, such as wind or solar, and discharges to displace dirty and expensive fossil-fuel generation to meet electricity demand.1Mallapragada D.S. Sepulveda N.A. Jenkins J.D. Long-run system value of battery energy storage in future grids with increasing wind and solar generation.Appl. Energy. 2020; 275: 115390https://doi.org/10.1016/j.apenergy.2020.115390Crossref Scopus (63) Google Scholar However, this ideal assumption is not always feasible in practice, particularly in deregulated power systems. Energy storage participates in electricity markets by submitting economic bids to earn revenue.2Brijs T. Belderbos A. Kessels K. Six D. Belmans R. Geth F. Energy storage participation in electricity markets.Advances in Energy Storage: Latest Developments from R&D to the Market. Wiley, 2022: 775-794https://doi.org/10.1002/9781119239390.ch35Crossref Scopus (0) Google Scholar Whether a storage unit charges or discharges at a specific time is not directly based on the system cost or carbon emissions but instead depends on market clearing, which is influenced by the storage’s bid prices, bids submitted by other participants, and the system’s demand and renewable share. Therefore, the storage’s market participation options and the market design play critical roles in determining whether storage can facilitate cost-effective decarbonization in future power systems. Previous studies examining storage integration in power systems have largely neglected the market participation aspect, assuming vertically integrated systems where system operators have direct control over storage. These studies have concluded that storage investments reduce the cost of electricity,3He G. Lin J. Sifuentes F. Liu X. Abhyankar N. Phadke A. Rapid cost decrease of renewables and storage accelerates the decarbonization of China’s power system.Nat. Commun. 2020; 11: 2486https://doi.org/10.1038/s41467-020-16184-xCrossref PubMed Scopus (137) Google Scholar,4Sepulveda N.A. Jenkins J.D. Edington A. Mallapragada D.S. Lester R.K. The design space for long-duration energy storage in decarbonized power systems.Nat. Energy. 2021; 6: 506-516https://doi.org/10.1038/s41560-021-00796-8Crossref Scopus (150) Google Scholar,5Schmidt O. Hawkes A. Gambhir A. Staffell I. The future cost of electrical energy storage based on experience rates.Nat. Energy. 2017; 2: 17110https://doi.org/10.1038/nenergy.2017.110Crossref Google Scholar,6Parag Y. Sovacool B.K. Electricity market design for the prosumer era.Nat. Energy. 2016; 1: 16032https://doi.org/10.1038/nenergy.2016.32Crossref Google Scholar,7Kittner N. Lill F. Kammen D.M. Energy storage deployment and innovation for the clean energy transition.Nat. Energy. 2017; 2: 17125https://doi.org/10.1038/nenergy.2017.125Crossref Scopus (535) Google Scholar,8Comello S. Reichelstein S. The emergence of cost effective battery storage.Nat. Commun. 2019; 10: 2038https://doi.org/10.1038/s41467-019-09988-zCrossref PubMed Scopus (106) Google Scholar,9Hanna R. Victor D.G. Marking the decarbonization revolutions.Nat. Energy. 2021; 6: 568-571https://doi.org/10.1038/s41560-021-00854-1Crossref Scopus (23) Google Scholar while the impact on carbon emissions is mixed and largely depends on the system resource mix.10Arbabzadeh M. Sioshansi R. Johnson J.X. Keoleian G.A. The role of energy storage in deep decarbonization of electricity production.Nat. Commun. 2019; 10: 3413https://doi.org/10.1038/s41467-019-11161-5Crossref PubMed Scopus (155) Google Scholar,11De Sisternes F.J. Jenkins J.D. Botterud A. The value of energy storage in decarbonizing the electricity sector.Appl. Energy. 2016; 175: 368-379https://doi.org/10.1016/j.apenergy.2016.05.014Crossref Scopus (251) Google Scholar,12Bistline J.E.T. Blanford G.J. Impact of carbon dioxide removal technologies on deep decarbonization of the electric power sector.Nat. Commun. 2021; 12: 3732https://doi.org/10.1038/s41467-021-23554-6Crossref PubMed Scopus (39) Google Scholar,13Zhang S. Chen W. Assessing the energy transition in china towards carbon neutrality with a probabilistic framework.Nat. Commun. 2022; 13: 87https://doi.org/10.1038/s41467-021-27671-0Crossref PubMed Scopus (82) Google Scholar,14Victoria M. Zhu K. Brown T. Andresen G.B. Greiner M. Early decarbonisation of the European energy system pays off.Nat. Commun. 2020; 11: 6223https://doi.org/10.1038/s41467-020-20015-4Crossref PubMed Scopus (75) Google Scholar,15Lin Y. Johnson J.X. Mathieu J.L. Emissions impacts of using energy storage for power system reserves.Appl. Energy. 2016; 168: 444-456https://doi.org/10.1016/j.apenergy.2016.01.061Crossref Scopus (51) Google Scholar,16Hittinger E.S. Azevedo I.M. Bulk energy storage increases united states electricity system emissions.Environ. Sci. Technol. 2015; 49: 3203-3210https://doi.org/10.1021/es505027pCrossref PubMed Scopus (76) Google Scholar,17Craig M.T. Jaramillo P. Hodge B.-M. Carbon dioxide emissions effects of grid-scale electricity storage in a decarbonizing power system.Environ. Res. Lett. 2018; 13: 014004https://doi.org/10.1088/1748-9326/aa9a78Crossref Scopus (35) Google Scholar However, these results may be too optimistic as they overlook the complexity introduced by market participation. Another group of studies has focused on storage’s participation in electricity markets, proposing various methods to control storage and design bids to optimally maximize market revenue.18Olsen D.J. Kirschen D.S. Profitable emissions-reducing energy storage.IEEE Trans. Power Syst. 2019; 35: 1509-1519https://doi.org/10.1109/TPWRS.2019.2942549Crossref Scopus (13) Google Scholar,19Davies D.M. Verde M.G. Mnyshenko O. Chen Y.R. Rajeev R. Meng Y.S. Elliott G. Combined economic and technological evaluation of battery energy storage for grid applications.Nat. Energy. 2019; 4: 42-50https://doi.org/10.1038/s41560-018-0290-1Crossref Scopus (174) Google Scholar,20Krishnamurthy D. Uckun C. Zhou Z. Thimmapuram P.R. Botterud A. Energy storage arbitrage under day-ahead and real-time price uncertainty.IEEE Trans. Power Syst. 2017; 33: 84-93https://doi.org/10.1109/TPWRS.2017.2685347Crossref Google Scholar,21Xu B. Zhao J. Zheng T. Litvinov E. Kirschen D.S. Factoring the cycle aging cost of batteries participating in electricity markets.IEEE Trans. Power Syst. 2017; 33: 2248-2259https://doi.org/10.1109/TPWRS.2017.2733339Crossref Scopus (182) Google Scholar,22He G. Chen Q. Moutis P. Kar S. Whitacre J.F. An intertemporal decision framework for electrochemical energy storage management.Nat. Energy. 2018; 3: 404-412https://doi.org/10.1038/s41560-018-0129-9Crossref Scopus (47) Google Scholar,23Arciniegas L.M. Hittinger E. Tradeoffs between revenue and emissions in energy storage operation.Energy. 2018; 143: 1-11https://doi.org/10.1016/j.energy.2017.10.123Crossref Scopus (29) Google Scholar,24Zheng N. Jaworski J. Xu B. Arbitraging variable efficiency energy storage using analytical stochastic dynamic programming.IEEE Trans. Power Syst. 2022; 37: 4785-4795https://doi.org/10.1109/TPWRS.2022.3154353Crossref Scopus (1) Google Scholar However, most of these studies have not considered how storage alters system operation, as they assume that storage is an emerging participant and will not impact market clearing prices, i.e., price takers. A few studies have attempted to analyze the impact of storage participation on market clearing results but have only focused on equilibrium analysis with small storage installations.25Awad A.S.A. Fuller J.D. El-Fouly T.H.M. Salama M.M.A. Impact of energy storage systems on electricity market equilibrium.IEEE Trans. Sustain. Energy. 2014; 5: 875-885https://doi.org/10.1109/TSTE.2014.2309661Crossref Scopus (81) Google Scholar,26Chen D. Jing Z. An improved market mechanism for energy storage based on flexible state of energy.CSEE J. Power Energy Syst. 2020; 8: 838-848https://doi.org/10.17775/CSEEJPES.2020.02980Crossref Scopus (5) Google Scholar,27Pandžić K. Pavić I. Andročec I. Pandžić H. Optimal battery storage participation in European energy and reserves markets.Energies. 2020; 13: 6629https://doi.org/10.3390/en13246629Crossref Scopus (13) Google Scholar,28Bhattacharjee S. Sioshansi R. Zareipour H. Energy storage participation in wholesale markets: the impact of state-of-energy management.IEEE Open J. Power Energy. 2022; 9: 173-182https://doi.org/10.1109/OAJPE.2022.3174523Crossref Scopus (1) Google Scholar,29Gu W. Sioshansi R. Market equilibria with energy storage as flexibility resources.IEEE Open J. Power Energy. 2022; 9: 584-597https://doi.org/10.1109/OAJPE.2022.3217973Crossref Scopus (0) Google Scholar,30Wang Y. Dvorkin Y. Fernandez-Blanco R. Xu B. Qiu T. Kirschen D.S. Look-ahead bidding strategy for energy storage.IEEE Trans. Sustain. Energy. 2017; 8: 1106-1117https://doi.org/10.1109/TSTE.2017.2656800Crossref Scopus (81) Google Scholar A recent study has investigated the market participation of virtual power plants using a multiscale model.31Gao X. Knueven B. Siirola J.D. Miller D.C. Dowling A.W. Multiscale simulation of integrated energy system and electricity market interactions.Appl. Energy. 2022; 316: 119017https://doi.org/10.1016/j.apenergy.2022.119017Crossref Scopus (8) Google Scholar The results highlight the importance of modeling bidding processes, which provides different results than price taker or vertically integrated studies. Understanding the impact of increasing storage participants in electricity markets on system cost and emissions is critical for guiding future market designs and regulatory incentives, especially given the rapid deployment of energy storage worldwide, which is driven by policy incentives and decreasing investment costs. Organized electricity markets in North America and Europe have allowed storage to participate and submit charge and discharge bids.32Federal Energy Regulatory CommissionElectric storage participation in markets operated by regional transmission organizations and independent system operators.2019https://www.federalregister.gov/documents/2019/05/23/2019-10742/electric-storage-participation-in-markets-operated-by-regional-transmission-organizations-andGoogle Scholar,33Sakti A. Botterud A. O’Sullivan F. Review of wholesale markets and regulations for advanced energy storage services in the united states: current status and path forward.Energy Policy. 2018; 120: 569-579https://doi.org/10.1016/j.enpol.2018.06.001Crossref Scopus (34) Google Scholar California is a leader in storage deployments, with total storage capacity participating in electricity markets surging from around 200 MW in 2020 to over 4,000 MW in 2022, accounting for 10% of California’s total electricity demand.34Murray C. California ISO adopts energy storage-friendly market reforms. Energy storage news.Energy Storage News. 2022; https://www.energy-storage.news/california-iso-caiso-adopts-energy-storage-friendly-market-reforms/Google Scholar As storage capacity increases, arbitrage in wholesale markets has replaced frequency regulation as the main market service for storage in California.35US Energy Information AssociationForm EIA-860 detailed data with previous form data (EIA-860A/860B).2022https://www.eia.gov/electricity/data/eia860/Google Scholar Similar trends in storage deployments and market choices are observed in other nations and regions, including Texas, Australia, and Germany.36Shan R. Reagan J. Castellanos S. Kurtz S. Kittner N. Evaluating emerging long-duration energy storage technologies.Renew. Sustain. Energy Rev. 2022; 159: 112240https://doi.org/10.1016/j.rser.2022.112240Crossref Scopus (25) Google Scholar This study aims to bridge the gap between storage participation bidding models and system-level cost and emission analysis. We propose an agent-based two-stage market model that employs innovative algorithmic designs to provide a more realistic and comprehensive analysis of storage’s impact on system cost and carbon emissions. The model includes accurate technical market-clearing and storage operation models, which yield more precise results. Our study investigates storage market participation perspectives that have been often overlooked by previous storage integration studies, such as bidding strategies and the choice of day-ahead or real-time markets. We perform a comprehensive analysis to understand how private incentives align with social welfare, including system cost and carbon emissions. This innovative setting enables us to uncover critical insights into the challenges and opportunities associated with storage market participation in cost-efficient decarbonization. We consider a typical two-stage electricity pool market clearing model, also called centralized market clearing, consisting of day-ahead and real-time markets, which is commonly used in North America37Kirschen D.S. Strbac G. Fundamentals of Power System Economics. John Wiley & Sons, 2018Google Scholar as shown in Figure 1. This wholesale electricity market model aims to match supply with demand economically while also fulfilling operating reserve provision requirements. One day before the operating day, the system operator collects predicted demand and wind generation data from participants38ISO New EnglandISO new England day-ahead enhancements.2018https://www.iso-ne.com/static-assets/documents/2018/12/20181205-da-enhancements-tech-session-1.pdfGoogle Scholar and then clears the day-ahead market by solving a unit-commitment problem to clear the market and to determine the start-up and shut-down schedule of thermal generators. The unit-commitment problem is expressed as follows:min∑t=1TCmgt+Cnut+Csyt+Ceptd(Equation 1) subject to unit commitment constraints, reserve requirements, generator operational constraints, and storage operational constraints, where Cm(gt), Cn(ut), Cm(yt) are generator operational cost, no-load cost, and start-up cost, respectively. Ce(ptd) is the marginal physical cost for discharge, considering degradation. T indicates the total time periods in a day. During the operating day, the system operator solves an economic dispatch problem (Equation 2) shortly before each time period (t) by utilizing the most recent system information and then updates the dispatch of generators accordingly:minCm(gt)+Btd·ptd−Btc·ptc(Equation 2) subject to generator operational constraints and storage operational constraints, where Btd and Btc indicate storage discharging and charging bids, respectively; and ptd and ptc are storage discharging energy and charging energy. To access detailed mathematical models of the day-ahead and real-time markets, please refer to supplemental information. We consider three participation options for storage over the proposed two-stage market-clearing model.1Day-ahead (DA) participation: storage participates in day-ahead markets by bidding its physical parameters. The system operator schedules storage in the day-ahead unit commitment with other generators over a 24-h horizon. Storage does not participate in real-time markets and does not respond to real-time price signals. Storage revenue in day-ahead markets is cleared using day-ahead prices.2Real-time (RT) participation: storage submits separate charge and discharge bids for each market period to participate in real-time markets. These bids are designed by solving a profit-maximization problem, as outlined in supplemental information. For each time period, the storage submits a charging bid, indicating the price below which it is willing to charge, and a discharging bid, indicating the price above which it is willing to discharge. We assume that the storage uses day-ahead price forecasts to design bids for the real-time markets. The system operator clears the storage bids, along with those of other generators, in real-time markets. The storage’s revenue in the real-time markets is cleared using real-time prices.3Day-ahead and real-time (DA + RT) participation: storage participates in both day-ahead and real-time markets. One day prior to the operating day, the storage participates in day-ahead markets following the rules of DA participation. During the operating day, the storage uses the published day-ahead price signal to design bids to update their buy (charge) and sell (discharge) positions in real-time markets (please note that DA + RT participation is not a bidding strategy based on stochastic programming. While participating in the DA, storage does not withhold charging/discharging capacity through bidding, which means they can buy back or sells more in real-time markets). The simulations are conducted using the independent system operator (ISO) New England test system.39Krishnamurthy D. Li W. Tesfatsion L. An 8-zone test system based on iso new England data: development and application.IEEE Trans. Power Syst. 2015; 31: 234-246https://doi.org/10.1109/TPWRS.2015.2399171Crossref Scopus (47) Google Scholar The system demand varies from 9 to 17 GW, with an average of 13 GW. The system has 76 generators with a total capacity of 23.1 GW. The generation mix (without renewables) includes: natural gas, 10.63 GW (46% of the total, cost range, $22.2–$400/MWh); nuclear, 4.66 GW (20.2%, cost range $5–$11/MWh); coal, 2.40 GW (10.4%, cost range $18.1–$20/MWh); and oil 5.4 GW (23.4%, cost range $54–$350/MWh). We select five representative demand and wind profiles for our study using a K-means approach, as explained in supplemental information, with the average wind capacity factor of 0.4. To model the impact of storage with growing renewable capacity, we scale the wind-generation capacity into three cases according to its maximum power capacity: low, medium, and high, representing 50% (6.5 GW), 100% (13 GW), and 200% (26 GW) average system demand, respectively. For each demand and wind scenario, we employ a Monte Carlo method to generate five real-time scenarios to consider real-time wind fluctuations, in which the average mean absolute error (MAE) of wind fluctuations is 11.53%. Thus, each demand-wind scenario corresponds to five real-time realizations, providing a total number of 25 scenarios for each wind-capacity case. We analyze the impact of integrating energy storage ranging from 1 MW to 5,000 MW into energy systems. All storage units considered in our analysis are 4-h batteries with a one-way charging/discharging efficiency of 90%, which is consistent with typical power systems’ resource-adequacy requirements.40California ISOCalifornia ISO energy storage enhancements issue paper.2021http://www.caiso.com/InitiativeDocuments/IssuePaper-EnergyStorageEnhancements.pdfGoogle Scholar We begin by examining the impact of storage on the suppliers’ cost of electricity generation and carbon emissions from a social welfare perspective of the power system operator. We simulate various storage participation options and compare the results as storage capacity increases. Figures 2A–2C display the average fuel costs of electricity generation, and Figures 2D–2F display the average carbon emissions across the three wind-penetration levels considered. Results show that storage participating in real-time markets can more effectively reduce system operating costs than in day-ahead markets. However, we also found that in the medium-wind case, increasing storage capacity can result in higher system costs in day-ahead markets. This is because storage displaces committed generators in the day-ahead market, leading to more frequent dispatch of expensive peaker generators in real-time to ensure generation/demand balance, which in turn causes price spikes and generation cost increments. Despite this, storage participation in day-ahead unit commitment can still lead to lower carbon emissions by displacing more thermal generators. Real-time markets provide the opposite effect of day-ahead markets over storage participation: Real-time markets are more effective in reducing generation costs while not significantly reducing carbon emissions. Storage shows a clear saturation effect in cost reduction, in which more storage capacity beyond a certain level—depending on the wind penetration—no longer reduces the system cost. In the low-wind scenario, only 1 GW of storage capacity is needed to achieve the lowest generation cost; in the medium- and high-wind scenarios, the lowest-cost storage capacity increases to 2 and 3.75 GW, respectively. Similar to the day-ahead case, we find that costs increase with high storage capacity, but the driving cause is different. The primary reason is that the cycle efficiency loss of storage increases the overall energy demand and outweighs the savings from energy shifting as storage capacity increases. Another cause is that storage bids may mismatch with demand peaks and valleys because the real-time prices become more deviated than day-ahead predictions. The market-clearing results may lead storage to charge during high demand and discharge during low demand, which drives up the cost of electricity. As storage capacity increases, participating in both day-ahead and real-time markets outperforms participating in day-ahead markets only regarding generation cost and carbon emission. The DA + RT cases fall between DA and RT in cost reduction but outperform DA and RT in carbon-emission reduction. In DA + RT participation, fewer thermal generators are committed in day-ahead unit commitment, while the storage still has the flexibility to buy back its day-ahead position in real-time and reduce the generation from peaker units compared with DA cases, leading to the overall lowest carbon emission. We also notice that storage participants always produce higher carbon emissions under the low-wind scenario than those having no storage in all participation options. This indicates that storage is more often charged from fossil-fuel generators than wind generation since renewable accommodation is very close to 100%. Hence the higher the storage capacity, the higher the emission due to storage’s charging/discharging efficiency loss. This section takes the perspective of storage participants to investigate the arbitrage profits under different market participation options. Figure 3 shows that storage profitability diminishes quickly as storage capacity increases. The per-unit storage profit in DA decreases at a steadier rate, which dropped to below $15 MWh per day at similar storage capacities in all three wind penetrations, while the storage profit in RT and DA + RT starts higher but reduces more quickly and even drops to negative. The results show that DA participation offers lower profits at low storage capacity, but that profits become more stable as storage capacity increases. In contrast, RT and DA + RT options provide high profits at low storage capacity, but the profit potential quickly diminishes as storage capacity increases, particularly in the low-wind case. The profitability of storage is closely correlated with price variability. In day-ahead markets, the combined effect of wind plus storage causes fewer baseload generators to be committed in unit commitments, while more expensive but also more flexible generators are kept to mitigate system variability. This helps to maintain price volatility, resulting in stable profitability. Figure 3 shows that, regardless of wind capacity, storage profits are always greater than zero and higher than the profits of RT or DA + RT models under large storage capacities. Conversely, for small storage capacity, the profits of DA participation are only half of those of RT participation because of its inability to capture large price volatility in real-time markets. The profits of RT participation under small storage capacity are much higher than those of DA participation due to higher price volatility, especially in the medium- and high-wind scenarios. However, the profits quickly diminish and may become negative as the storage capacity increases. This is because the price predictions used to generate bids become more deviated from the actual market prices due to storage’s participation. By comparing Figures 3 and 2, we notice that there are close saturation points in storage profits and generation costs in real-time markets: 1 GW storage for low wind, 2 GW storage for medium wind, and 3.5 GW storage for high wind, respectively. This shows that, despite errors in bid designs, storage’s profit objective aligns with social welfare in reducing generation costs. DA + RT participation is less profitable than RT participation under all wind-penetration scenarios. Our study assumes storage to be truthful bidders in day-ahead markets by submitting only physical costs and parameters. As a result, storage often clears a portion of its capacity in day-ahead markets and must buy those capacities back in real-time markets to arbitrage more volatile real-time prices. This approach lowers the storage profit as storage did not design DA bids strategically by considering real-time arbitrage opportunities. Therefore, our study suggests that storage participants may prefer to participate in real-time markets only to earn maximal profits and tend to avoid being scheduled in day-ahead markets. This conclusion aligns with observations from the recent California market, where storage participants bid unreasonably high prices in day-ahead markets with the purpose of not being cleared.41California ISOCAISO 2021 annual report on market issues and performance.2021http://www.caiso.com/Documents/2021-Annual-Report-on-Market-Issues-Performance.pdfGoogle Scholar We now examine the impact of energy storage on the cost of electricity and carbon emissions from the perspective of consumers. Based on our previous findings on the disparate effects of day-ahead and real-time market participation options for energy storage, we use Pareto frontiers to analyze how different participation options of energy storage affect energy affordability, measured as average consumer payment (A brief introduction of electricity market pricing principle is provided in supplemental information) calculated as the demand-weighted average of market clearing prices, and sustainability, measured as average carbon emissions. Note that consumer payment in this study refers only to payment from the wholesale market, which takes up around 30% of the consumer’s utility bill, while the rest arose from distribution grid-management and