Grid infrastructure and renewables integration for singapore energy transition

Abstract

Considering rising environmental concerns and the energy transition towards sustainable energy, Singapore’s power sector stands at a crucial juncture. This study explores the integration of grid infrastructure with both generated and imported renewable energy (RE) sources as a strategic pathway for the city-state’s energy transition to reach net-zero carbon emissions by 2050. Employing a combination of simulation modeling and data analysis for energy trading and advanced energy management technologies, we examine the current and new grid infrastructure’s capacity to assimilate RE sources, particularly solar photovoltaic and energy storage systems. The findings reveal that with strategic upgrades and smart grid technologies; Singapore’s grid can efficiently manage the variability and intermittency of RE sources. This integration is pivotal in achieving a higher penetration of renewables, as well as contributing significantly to Singapore’s commitment to the Paris Agreement and sustainable development goals. While the Singapore’s power system has links to the Malay Peninsula, the planned ASEAN regional interconnection might alter the grid operation in Singapore and possibly make Singapore a new green energy hub. The study also highlights the key challenges and opportunities associated with cross-border energy trade with ASEAN countries, including the need for harmonized regulatory frameworks and incentives to foster public–private partnerships. The insights from this study could guide policymakers, industry stakeholders, and researchers, offering a roadmap for a sustainable energy transition in Singapore towards meeting its 2050 carbon emission goals.

Introduction

Recent changes in climatic conditions of extreme weather events like heatwaves in the northern hemisphere and rainfall in Europe and Asia. These events renewed the focus on holding global warming to 1.5 degree Celsius by adopting energy transition in the Paris Agreement, 2015^1,2. These extreme weather events result in transition into vulnerable nations like Singapore, which face disproportionate climate risks due to their geography and developmental constraints.

Singapore (SG) is experiencing the global warming effect with rise in the temperature by 0.25 °C per decade from 1948 to 2015, while 2016 and 2019 were recorded as the hottest years. As a low-lying city-state, the island nation is vulnerable to rising sea levels due to extreme weather events. Recently, a third national climate change study was conducted by the Centre for Climate Research Singapore (CCRS)³, and a rise in sea level of 0.25 m to 1.15 m by 2100 and up to around 2 m by 2150 with an increase in global temperature of 0.55 °C per decade until 2100 is predicted. Also, the study projected an increase in rainfall by about 6% to 92% in April and May by the end of the century⁴. Furthermore, the country uses fossil fuel as the major generating sources to meet the load demand due to its limited land constraints. The peak CO₂ emission is estimated to be 65 MtCO₂e around 2030 and its value will be reduced by half in the second half of the century to 33 MtCO₂e by 2050 to achieve net zero emissions⁵. In 2019, the SG government announced the Southeast Asia first carbon tax, initially set as SGD $5 per tonne of GHG emissions (tCO₂e) from 2019 to 2023, with plans to rise to SGD$50–80 by 2030.

Despite early leadership, SG faces a complex trilemma: how to reconcile energy security, affordability, and sustainability in a fossil-reliant, land-scarce city-state. Domestic renewable potential is limited, and decarbonization depends heavily on regional electricity imports, alternative fuels like hydrogen, and the modernization of grid infrastructure. However, the scientific and strategic integration of these components remains underexplored in existing literature. Thus, this study explores Singapore’s integrated energy transition roadmap, with a focus on innovations like regional grid connectivity, hydrogen deployment, and digital energy systems.

Background

Singapore aims to align its development goals with energy affordability by importing electricity from neighboring ASEAN countries and adopting clean energy sources, while safeguarding energy security⁶. As pledged at the UN Climate Summit in 2019, SG has been at the forefront among the ASEAN countries to reduce greenhouse gas emissions intensively by 2050. By being the first country in the region to accelerate decarbonization, SG introduced a carbon tax in 2019. The value was set at SGD$5 (USD 3.67) per tonne of carbon dioxide emitted and its value is SGD$25 per tonne in 2024—2025 with a view of reaching SGD$50 to SGD$80 per tonne by 2030⁷. However, the challenge for the island nation going forward is the space constraints to install solar panels within the limited space.

Over the years, solar deployment has grown significantly, surpassing 1 GWp of the grid-connected installed capacity in the first half of 2023. Figure 1 gives the installed capacity of solar power in the city-state over the last 6.5 years. The island nation is on track to meet 1.5 GWp by 2025, and at least 2 GWp by 2030. The major driving force for such a growth of solar deployment is the private sector which accounts for 63.5% (or 638.8 MWp) of the total installed capacity. Town Councils and Public Housing Common Services also made substantial contributions of 25.9% (or 260.4 MWp). On the other hand, the Public Service Agencies and Residential installations contributed 6.6% (or 66.0 MWp) and 4.0% (or 40.4 MWp), respectively⁸. Solar power is a dominant source of Singapore’s renewable energy, however, due to its land scarcity, there is a constraint in deploying it on a larger scale.

Fig. 1

Solar power installed capacity in Singapore.

On the other hand, to meet the net-zero carbon emission, the Energy Market Authority (EMA) of SG has created a committee called Energy 2050. Some of their strategies of achieving zero emissions include importing of electricity from nearby ASEAN countries, the usage of low-carbon hydrogen instead of natural gas, use of alternative ways to generate electricity and other strategies⁹. Considering these strategies and during transition to low-carbon hydrogen, the Ministry of Trade and Industry (MTI) recently announced that the EMA plans to set up a central gas company in 2024 called ‘Gasco’. It will centralize the procurement of natural gas from diverse sources and its supply to the power sector by aggregating the gas demands from the ‘Gencos’. In doing so, Gasco will have a stronger negotiating level, and be in a better position to negotiate for more favorable contracting terms and reduce concentration risk. The ministry added that natural gas plays a significant role in supplying 95% of the country’s electricity demand¹⁰.

Energy security in SG energy transitions

Ensuring secure, stable, and diversified energy supply is a top priority in Singapore’s energy transition. Given its limited domestic energy resources, Singapore has prioritized resilient infrastructure, diverse sourcing, and regional cooperation to strengthen energy security. A major component of the current energy mix is natural gas, which is imported primarily from Malaysia and Indonesia. Additionally, liquified natural gas (LNG) processing facilities on Jurong Island, which are managed by Singapore LNG Corporation (SLNG), constitute the third source. LNG is transported via authorized importers, including Shell Eastern Trading, Pavilion Energy SG, Sembcorp Fuels, and ExxonMobil LNG Asia Pacific, from the United States, Qatar, and Australia¹¹, respectively. A second LNG terminal for SLNG will be operational by the end of this decade. The second terminal, in contrast to its Jurong Island-based counterpart, consists of a floating storage and regasification facility (FSRU). The FSRU will be a vessel equipped with onboard facilities for converting LNG into gaseous form, which is then routed to the SG gas infrastructure onshore. Although FSRU will be berthed at Jurong Island, it is capable of being reassigned to any other location within SG.

It is expected that more than half of the power demand will depend on natural gas by 2035. Power generation based on natural gas accounts for 40% of SG’s carbon emissions. This can be reduced in the future through the retrofitting of gas turbines with repowering capabilities and a minimum thirty percent hydrogen fuel blend¹². Additionally, SG will import RE from multiple nations in 2035, which will account for 30% of the energy composition. One gigawatt of hydropower may be delivered to SG by 2032 via 700-km-long submarine cables from the Malaysian state of Sarawak. A comparable method of power transmission through submarine cables is effectively implemented in Europe, spanning over 720 km in the North Sea, connecting Norway and Britain¹³.

SG intends to import 1.4 GW of offshore wind energy from Vietnam via 1000-km-long subsea cables by 2033. A 2-GW solar and battery project on the Riau Islands of Indonesia aims to provide electricity to SG prior to 2030. Additionally, SG will receive 1 GW of power from Cambodia, sourced from a combination of solar, wind, and hydropower, transmitted via undersea cables spanning 1000 km. At present, 100 MW of hydropower is exported to SG via the Lao PDR-Thailand-Malaysia-Singapore Power Integration Project, which utilizes pre-existing interconnections for cross-border electricity trading. All four nations were engaged in discussions regarding strategies to bolster their cross-border trade capabilities in all directions, thereby contributing to the advancement of the ASEAN regional power grid. This opens an opportunity for interconnection of Southeast Asian countries and contributes towards the ASEAN grid vision. The EMA has received more than 20 proposals from six countries like Australia, Cambodia, Indonesia, Laos, Malaysia, and Thailand to export solar, wind, hydro and geothermal power to SG to support the decarbonization of the island nation¹⁴. Figure 2 summaries the power and natural gas import at present and in the near future to SG from other countries^15,16.

Fig. 2

Geographical map of power and natural gas import to SG from other ASEAN countries.

Consequently, SG and the United States have concluded the preliminary stage of their study investigating the social and economic advantages of regional energy connectivity. Emission reduction decreased capital and production costs, increased resource sufficiency, and a resilient power supply are all aspects of the study. The United States and SG intend to commence the subsequent stage of the investigation, which will center on the framework of governance and financing for the execution of cross-border energy trading initiatives¹⁷. Figure 3 depicts the geographical view of ASEAN power grid interconnection. It is seen that the ASEAN regional grid is divided into three regions, namely Northern, Southern, and Eastern regions, and there are 16 interconnection projects, among which a few are already in operation. Some are ongoing, and the remaining will be established in the future. The status of interconnection and further details are given in the Heads of ASEAN Power Utilities/Authorities (HAPUA) secretariat¹⁸. The regional grid has five major working groups. SG being one among them, is responsible for developing the software solution for “Distribution Power and Reliability & Quality” and serves as a chair in harmonization with Myanmar. The objective is to discuss the impact and smoothing of the integration of distributed energy resources (DERs) and to support HAPUA in achieving the goal of 23% renewable energy (RE) integration by 2025.

Fig. 3

Geographical map of ASEAN regional power grid interconnection¹⁹.

Energy affordability through strategic planning and market design

To ensure energy remains affordable while pursuing decarbonization, Singapore has adopted a multi-scenario approach with long-term planning, technological innovation, and market-based mechanisms. EMA has analyzed and published the Energy 2050 Committee report, laying down the possible net-zero ways for the power sector through the energy mix as given in Fig. 4 from 2023 until 2050. The committee has examined the development of grid infrastructure, supply technologies, and power demand trends in the context of critical uncertainties and predetermined trends that pave the way for the three scenarios: Clean Energy Renaissance, Climate Action Bloc, and Emergent Technology Trailblazer.

Fig. 4

Energy mix to achieve net-zero carbon emissions by 2050.

The Clean Energy Renaissance is characterized by fast advancements in energy and digital technology, bolstered by robust international collaboration. In 2050, SG may accomplish a diverse supply mix primarily through the import of energy and low-carbon hydrogen as given in Fig. 4(c). Importing power is a safe and inexpensive supply alternative due to a diverse portfolio of sources and a local backup capacity. Low-carbon hydrogen is becoming more competitive in price due to global advances. Two other important domestic renewable generating sources in SG are geothermal and solar. Despite the rapid expansion of DERs, SG has developed a world-class intelligent grid that can keep the system stable and reliable. By utilizing smart technologies like artificial intelligence (AI) and machine learning (ML), the grid can optimize the system locally, taking advantage of the DER’s flexibility, and ultimately decreasing grid capacity and costs. Grid planning and operations are made more efficient by actively managing the rise of total power demand and smoothing out the erratic demand profile. More and more consumers are using cutting-edge energy and digital technology to gauge and control their usage based on their own goals and requirements.

As progress in technology stalls or comes to a near halt in the Climate Action Bloc, nations unite. To meet its electrical demands in 2050, SG would have to depend majorly on power imports from other countries as illustrated in Fig. 4(d). All the region’s demands may be met, and economic advantages can be distributed through a larger regional grid and trading platform. The expensive cost of hydrogen means that it only generates a small portion of Singapore’s electricity. Even though SG uses foreign carbon credits to offset its carbon emissions, it nevertheless uses a small quantity of natural gas. Despite having to balance and manage a diverse variety of DERs, Singapore’s grid is still one of the most dependable and robust in the world. Powerful grid planning and asset optimization are made possible with the use of digital twins and other modeling and simulation technologies. Asset cost optimization reduces the need to build back-up capacity to deal with demand and supply uncertainties. Large end users frequently employ micro-grids to increase system resilience and reap financial benefits. By doing optimization within the micro-grid, it is possible to influence the demand profile at the system level. There can be significant savings for the system when micro-grids are completely self-sufficient and do not need backup from the utility grid.

Technological advancement experiences a deceleration, yet it eventually resumes in the early 2040 s in Emergent Technology Trailblazer. Singapore’s primary fuel source for electricity in 2050 will be low-carbon hydrogen, replacing natural gas as given in Fig. 4(e). Rising worldwide deployment and economies of scale are causing hydrogen prices to reduce steadily over time. As a result of the regional grid’s delayed expansion and links with possible RE sources, electricity imports play a role in the energy mix, but they only account for a small portion of the total. From its initial investments, SG can now diversify its supply mix by installing additional low-carbon options, such as nuclear power, and is prepared to build them up even more when they become more financially attractive. Incorporating AI into the grid has greatly improved its intelligence, allowing for more accurate demand and supply projections and, ultimately, more autonomy in making decisions that maximizes resource use. Leveraging AI-driven orchestration of distributed energy resources (DERs), grid operators can aggregate and dynamically control heterogeneous assets like a virtual power plant (VPP), enabling flexible, demand-responsive system operations at scale. End users at all levels actively participate in managing and optimizing demand via the implementation of efficient energy efficiency and conservation measures. To improve their business operations and reap the financial benefits of energy efficiency, large industrial and commercial users invest in digital solutions. Diversified energy pricing and access to several energy sources encourage smaller end users to take an active part in environmental initiatives.

Digitalization and smart grids

Singapore’s sustainability strategy emphasizes the integration of digital energy solutions to support its commitment to net-zero emissions. To this end, in addition to importing power from other countries for decarbonization, decentralized control, and grid intelligence could revolutionize today’s grid architecture. In a sense, developing a digital solution for the grid using AI and ML technologies could help aggregate DERs such as rooftop solar PV systems, and coordinate them intelligently and efficiently to support the energy mix. This can be achieved through the development of an advanced energy management system (EMS) that can coordinate the solar PV power producers, with the energy storage system (ESS) that acts as a flexible resource and plays a significant role in the energy transition.

Other recently evolved digital solutions like peer-to-peer (P2P) energy trading using blockchain technologies, and digital twins, can provide a high-fidelity to motivate more and more private sectors and residents in the island nation to take part in reaching the net-zero emissions. Furthermore, the development of vehicle-to-grid (V2G) and vehicle-to-prosumers (V2P) concepts can enable electric vehicle (EV) owners to participate in electricity trading and receive compensation for their services. This concept leverages the home and building EMS to dynamically respond to the price signals and adjust the electricity consumption during the peak hours and ultimately reduce the consumption cost.

Literature review

As global concern over climate change intensifies, the shift towards renewable energy particularly has become essential for achieving emission reduction targets²⁰. International climate policies, such as the Paris Agreement²¹, have set a clear course for transitioning to clean energy sources, motivating countries worldwide to adopt sustainable energy practices. As an island nation with rising energy demands, Singapore faces unique challenges in its energy transition, including limited land resources, variable climate conditions, and a high dependency on external energy sources. Against this backdrop, smart grid²² and digital twin²³ technologies are being increasingly implemented worldwide to improve grid stability amid renewable integration. Smart grids provide flexibility to energy systems, allowing for real-time data management and automated control to optimize energy supply and demand balance. These technologies demonstrate significant potential in addressing the grid fluctuations associated with intermittent renewable sources, such as solar and wind energy. Furthermore, the development of Distributed Energy Resource Management Systems (DERMS) and advanced Energy Management Systems (EMS) offers a robust solution for integrating decentralized renewable energy sources into the grid. Recognizing the importance of digital solutions, the Singaporean government has introduced policies and innovative initiatives aimed at advancing the local power market towards a Smart Grid 2.0²⁴. The government has implemented measures such as demand response and energy storage systems to enhance the resilience of the energy system and has begun exploring blockchain technology for peer-to-peer (P2P) energy trading²⁵. P2P trading not only invigorates the electricity market but also offers opportunities for residents and businesses to participate in green energy transactions, fostering energy diversity and sustainability at the local level²⁶.

Cross-border energy collaboration is a key component of Singapore’s renewable energy strategy. With the gradual development of the ASEAN regional power grid, Singapore can secure clean energy imports from neighboring countries through submarine cables and cross-border agreements, thereby strengthening its energy security²⁷. Additionally, Singapore is exploring alternative fuels, such as hydrogen²⁸, to further enhance the cleanliness and diversity of its energy system. Hydrogen, with its potential as a low-carbon fuel, could significantly transform Singapore’s energy structure. However, it remains economically uncompetitive due to high production and infrastructure costs and technically constrained by the need for retrofitting existing gas systems. Policy gaps such as the lack of targeted incentives, regulatory frameworks, and mechanisms to integrate hydrogen into distributed systems or VPP participation further hinder its scalability²⁹. In summary, Singapore’s energy transition strategy is propelled by global trends, technological advancements, and regional partnerships. By refining existing policies and integrating cutting-edge technologies, Singapore can navigate the challenges of energy restructuring, laying a solid foundation for the innovative solutions and strategies discussed in this study to achieve its net-zero emissions goals. An energy management model based on rolling optimization was developed in a microgrid system (MG) containing a wind turbine (WT), a photovoltaic (PV) cell, a diesel engine generator (DE), energy storage (ES), and loads³⁰. This model firstly identifies the reference scheduling plans for the next two days based on a long-term prediction of source-load power, then updates the dispatch plans by a rolling optimization with a 1-h time scale to minimize the operation cost. To handle the uncertainties in power generation and load demand, the authors proposed a rolling optimization model that manages the real-time energy of MGs³¹. This model minimizes the operation cost of an MG by cyclically optimizing the predicting and controlling levels and adjusting the power deviations between the two levels through the controllable loads. To cope with uncertainty in rolling optimization, Ding et al.³² explored a rolling optimization method with multi-timescales, which maximizes the economic benefits of MGs. The day-ahead dispatch is determined by stochastic optimization approach, and the power deviations are minimized by penalty functions. However, the above-mentioned rolling optimal scheduling has not considered the pricing of intermittency mitigation in renewable generation management.

Motivation and contribution

The primary motivation of this study is to analyze Singapore’s strategy for achieving net-zero carbon emissions, focusing on its transition from fossil fuels to renewable energy sources and its alignment with the Paris Agreement’s 1.5 °C target. The objectives of the study include evaluating the challenges and solutions associated with renewable energy integration in a land-constrained city-state, particularly the growing reliance on solar power. The study also seeks to assess Singapore’s collaborative energy import initiatives with neighboring ASEAN countries, exploring how imports of hydropower, wind, and solar energy will contribute to the country’s energy mix by 2035. Additionally, this research examines the role of advanced digital innovations, such as artificial intelligence, machine learning, and blockchain technology, in enhancing grid intelligence, facilitating demand response, and supporting peer-to-peer energy trading within Singapore’s smart grid framework. By examining these elements, the study aims to identify the potential of decentralized control and advanced energy management systems, alongside emerging technologies like vehicle-to-grid and vehicle-to-prosumer systems, in achieving a resilient, sustainable, and economically feasible path to net-zero emissions for Singapore.

The paper is structured as follows: Section“Methods” describes the methodology of the proposed work. Section“Results” explains the results and discussion of digital solutions such as Hierarchical Energy Management System (Hi-EMS) and P2P energy trading to support energy transition. Section“Challenges during transition” and Section“Grid emission factor” present the challenges faced during the energy transition to reach the net-zero emission by 2050 and Grid emission factor. Finally, the conclusion is summarized in the last section of this paper.

Methods

Proof-of-authority: powering private blockchains with reputation

Proof-of-Authority (PoA) is a distinctive consensus mechanism in the realm of blockchain technology, emphasizing the significance of reputation rather than computational prowess or token holdings. In Po networks, validators, akin to a board of directors with established credibility, are handpicked entities entrusted with the responsibility of validating transactions and appending them to the blockchain-distributed ledger. This approach diverges from Proof-of-Stake (PoS) and Proof-of-Work (PoW) by prioritizing efficiency, scalability, energy-friendliness, and governance³³.From a practical implementation perspective, PoA outshines PoS and PoW in multiple aspects. The efficiency of PoA is evident as it sidesteps the computationally intensive processes associated with mining or the need to lock away substantial token amounts. Validator selection in PoA demands minimal computing power, resulting in quicker transaction processing and reduced fees. This makes PoA particularly advantageous for permissioned blockchains within a coalition of participants where both performance and cost-effectiveness are critical. Moreover, the scalability of PoA networks is noteworthy, as a limited number of trusted validators ensures swift and efficient consensus, enabling high transaction volumes without compromising speed or security. Additionally, PoA stands out for its energy-friendliness, contributing to a greener alternative by minimizing energy consumption compared to PoW mining, which often leads to a significant environmental impact³⁴.

PoA consensus mechanism is chosen for the P2P energy trading model due to its efficiency and scalability, which are essential for managing the real-time energy transactions. In PoA, validators play a crucial role in maintaining the network integrity by authorizing transactions. Validators are selected based on the established criteria, which can include their historical performance in transaction accuracy, technical competence, and overall reliability in maintaining network stability. This selection process ensures that only reputable and capable nodes assume the validation role, enhancing the trustworthiness of the network.

Over time, validator reputation is managed through continuous performance assessments, where validators are monitored for adherence to network standards, low error rates, and active participation. Validators that fail to maintain these standards may lose their authority, while high-performing validators are recognized and retained. Additionally, to address risks associated with centralization or collusion, PoA may incorporate strategies such as periodic rotation of validator roles or multisignature verification, which distributes decision-making power across multiple validators. These measures mitigate the risks of monopolistic control or collusion among validators, preserving the decentralization and fairness necessary in P2P energy trading.

Double-auction mechanism for P2P

A continuous double-auction algorithm is incorporated into the blockchain network’s smart contract for P2P energy trading, which is deployed before market initiation and remains unalterable during market operation. At the end of each trading period, the smart contract automatically executes the double auction algorithm to determine the market clearing price and trading quantities of each prosumer. The buyers’bids are arranged in descending order:

$${\text{b}}_{{1}}> {\text{ b}}_{{2}}> \cdots> {\text{ b}}_{{\text{N}}}$$

(1)

where b₁, b₂, …, b_N are the prices of consumers participating in the trading. Similarly, sellers’ask prices are arranged in ascending order:

$${\text{a}}_{{1}} < {\text{ a}}_{{2}} < \cdots < {\text{ a}}_{{\text{M}}}$$

(2)

where a₁, a₂, …, a_N are the cost of prosumers participate in the trading. The prosumers can participate in intra-hour energy trading markets which operate at regular intervals, such as every 10–15 min, for bilateral energy sharing. Unlike wholesale markets, buyers and/or sellers in P2P energy trading can determine their preferred transaction quantity and price one market interval in advance. Consumers and prosumers are required to submit their buying and selling prices and quantities to the blockchain energy trading platform before the end of the trading period. Detailed information on P2P energy trading and the PoA mechanism is described in Veerasamy et al³⁵.

In the P2P energy market, the double-auction mechanism serves as the primary trading structure, facilitating transparent and efficient energy transactions among prosumers. Bidding strategies in this system are designed to promote rational and fair participation. To prevent market manipulation, measures such as bid caps and auction transparency are enforced. Bid caps limit the extent of price fluctuations, while transparency requirements ensure that participants have access to accurate information on market conditions, enabling informed bidding decisions.

Another critical aspect is the impact of competition levels, particularly in smaller, localized energy markets with limited participants. In these markets, reduced competition can potentially lead to suboptimal auction outcomes, such as limited price discovery or increased price volatility. To address this, the system could implement minimum bid thresholds to encourage consistent demand and supply within the market. Additionally, incentives could be offered to attract more participants, thus fostering a competitive environment that enhances the auction’s efficiency and fairness. These considerations help maintain a balanced market structure and ensure that the double-auction mechanism functions effectively, even in constrained or low-participation scenarios. By incorporating these elements, the PoA and double-auction mechanisms within our P2P energy trading model are positioned to support a secure, fair, and resilient trading environment, promoting active participation and minimizing risks associated with validator centralization and market manipulation.

The P2P market model would promote local RE consumption, reducing reliance on centralized generation and fossil fuels, thereby supporting SG’s clean energy goals. Considering the practical application in the context of an energy market regulator like EMA, adopting a PoA-based P2P energy trading system over the distribution network can pave the way for integrating more renewable sources and reducing carbon footprints. The efficiency and scalability of PoA ensure rapid and secure transactions, while its energy-friendliness aligns with environmental goals. By leveraging PoA, the EMA could establish a transparent and accountable network of trusted validators, ensuring governance and facilitating the seamless exchange of energy between consumers and producers. This innovative approach holds the potential to transform SG’s energy landscape, fostering sustainability and embracing renewable sources to meet future energy needs while mitigating the environmental impact. In the future, one may expect energy trade between the communities or even with the countries via interconnected ties.

Intermittency mitigation using Hi-EMS

The Hi-EMS module aims to formulate a decision-making tool to generate control signal commands for optimization of BESS and DG operations. The goal is to achieve the coordination of all power units while mitigating solar intermittency and lowering the operation cost. The total operation cost considered includes the degradation cost of BESS, the start-up, and shutdown cost of DGs, the operation and maintenance cost of PV, and the trading cost with the main grid. To reduce the impact of predicted deviations in PV outputs, a bilevel rolling optimization framework was established for the EMS. According to the two-time scale rolling optimization framework described above, we develop the rolling optimization model for the EMS. Firstly, the 24-h ahead scheduling model is formulated as:

24-h ahead scheduling objective

$$\mathop {\min }\limits_{\substack{ U_{{{\text{DE,}}t}} ,I_{{{\text{DE,}}t}} ,M_{{{\text{DE,}}t}} ,U_{{{\text{ES,}}t}}^{{{\text{ch}}}} ,U_{{{\text{ES,}}t}}^{{{\text{dis}}}} ,U_{{{\text{GL,}}t}}^{{{\text{buy}}}} ,U_{{{\text{GL,}}t}}^{{{\text{sell}}}} \\ P_{{{\text{PV,}}t}} ,P_{{{\text{DE,}}t}} ,P_{{{\text{ES,}}t}}^{{{\text{ch}}}} ,P_{{{\text{ES,}}t}}^{{{\text{dis}}}} ,P_{{{\text{L,}}t}}^{{{\text{cut}}}} ,P_{{{\text{L}},t}}^{{{\text{tran}}}} ,P_{{{\text{GL,}}t}}^{{{\text{buy}}}} ,P_{{{\text{GL,}}t}}^{{{\text{sell}}}} } } \left( \begin{gathered} C_{{{\text{DE}}}}^{{{\text{on}}}} + C_{{{\text{DE}}}}^{{{\text{off}}}} + C_{{{\text{DE}}}}^{{{\text{fuel}}}} + C_{{{\text{DE}}}}^{{{\text{O}}\& {\text{M}}}} + \\ C_{{{\text{ES}}}}^{{{\text{O}}\& {\text{M}}}} + C_{{{\text{PV}}}}^{{{\text{O}}\& {\text{M}}}} + C_{{{\text{ES}}}}^{{{\text{loss}}}} + C_{{\text{L}}}^{{{\text{cut}}}} + C_{{{\text{GL}}}}^{{{\text{trade}}}} \\ \end{gathered} \right)$$

(3)

where:

$$C_{{{\text{DE}}}}^{{{\text{on}}}} = m_{{{\text{DE}}}}^{{{\text{on}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {I_{{{\text{DE,}}t}} } ,C_{{{\text{DE}}}}^{{{\text{off}}}} = m_{{{\text{DE}}}}^{{{\text{off}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {M_{{{\text{DE,}}t}} }$$

(4)

$$C_{{{\text{DE}}}}^{{{\text{fuel}}}} = m_{{{\text{DE}}}}^{{{\text{fuel}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {\left( {a_{{{\text{DE}}}} P_{{{\text{DE,}}t}} + b_{{{\text{DE}}}} P_{{{\text{DE}}}}^{{\text{r}}} U_{{{\text{DE,}}t}} } \right)} \Delta t$$

(5)

$$C_{{{\text{DE}}}}^{{{\text{O}}\& {\text{M}}}} = m_{{{\text{DE}}}}^{{{\text{O}}\& {\text{M}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {P_{{{\text{DE,}}t}} } \Delta t$$

(6)

$$C_{{{\text{ES}}}}^{{{\text{O}}\& {\text{M}}}} = m_{{{\text{ES}}}}^{{{\text{O}}\& {\text{M}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {\left( {P_{{{\text{ES,}}t}}^{{{\text{ch}}}} + P_{{{\text{ES,}}t}}^{{{\text{dis}}}} } \right)} \Delta t$$

(7)

$$C_{{{\text{PV}}}}^{{{\text{O}}\& {\text{M}}}} = m_{{{\text{PV}}}}^{{{\text{O}}\& {\text{M}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {P_{{{\text{PV,}}t}} } \Delta t$$

(8)

$$C_{{{\text{ES}}}}^{{{\text{loss}}}} = m_{{{\text{ES}}}}^{{{\text{loss}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {\left( {P_{{{\text{ES,}}t}}^{{{\text{ch}}}} + P_{{{\text{ES,}}t}}^{{{\text{dis}}}} } \right)} \Delta t$$

(9)

$$C_{{\text{L}}}^{{{\text{cut}}}} = m_{{\text{L}}}^{{{\text{cut}}}} \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {P_{{{\text{L,}}t}}^{{{\text{cut}}}} } \Delta t$$

(10)

$$C_{{{\text{GL}}}}^{{{\text{trade}}}} = \sum\limits_{t = 1}^{{N_{{\text{t}}} }} {\left( {pr_{{{\text{GL,}}t}}^{ * } P_{{{\text{GL,}}t}}^{{{\text{buy}}}} - \sigma \cdot pr_{{{\text{GL,}}t}}^{ * } P_{{{\text{GL,}}t}}^{{{\text{sell}}}} } \right)} \Delta t$$

(11)

Equation (3) is the objective function of the 24-h ahead scheduling model where Con DE/Coff DEis the startup/shutdown cost of the generator; Cfuel DE is the fuel cost of the generator; CO&M DE/CO&M ES/CO&M PV is the operation and maintenance cost of the generator/energy storage/PV; Closs ES is the loss cost (also as the degradation cost) of the energy storage; Ccut Lis the penalty cost of load shedding; and Ctrade GL is the cost for power trading in the grid-connected line.

Equations (4) - (11) denote the cost functions of the corresponding terms in (3). In (4), mon DE/moff DE is the cost coefficient for generator startup/shutdown; and I_DE,t/M_DE,t is the startup/shutdown state of the generator. t is the time index for the 24-h ahead scheduling, and N_t and Δt denote the number and period interval for the 24-h ahead scheduling. In (5), mfuel DE is the price of fuel; a_DE/b_DE is the fuel cost coefficient of the generator; Pr DE is the rated power of the generator; U_DE,t is the on/off state of the generator; and P_DE,t is the output power of the generator at period t. In (6)-(8), mO&M DE/mO&M ES/mO&M PV is the operation and maintenance cost coefficient of the generator/energy storage/PV; Pch ES,t/Pdis ES,t is the charge/discharge power of energy storage at t; and P_PV,t is the output of PV at t. In (9)-(10), mloss ES is the loss cost coefficient of the energy storage; mcut L is the penalty cost coefficient for load shedding; and Pcut L,t is the power of load shedding at t. In (11), Pbuy GL,t/Psell GL,t represents the buying/selling power of the grid-connected line at t, and pr* GL,t is the predicted buying price with the power grid at period t. σ is the proportion of the selling price to the buying price.

24-h ahead scheduling constraints

PV output constraints

$$P_{{{\text{PV,}}t}} = p_{t}^{ * }$$

(12)

Constraint (12) indicates that the PV output equals the predicted value p* t for each period.

Generator constraints

$$\sum\limits_{t = k}^{{k + N_{{{\text{DE}}}}^{{{\text{on,}}\min }} - 1}} {I_{{{\text{DE,}}t}} } \le 1,\sum\limits_{t = k}^{{k + N_{{{\text{DE}}}}^{{{\text{on,}}\min }} - 1}} {\left( {I_{{{\text{DE,}}t}} + M_{{{\text{DE,}}t}} } \right)} \le 1$$

(13)

$$\sum\limits_{t = k}^{{k + N_{{{\text{DE}}}}^{{{\text{off,}}\min }} - 1}} {M_{{{\text{DE,}}t}} } \le 1,\sum\limits_{t = k}^{{k + N_{{{\text{DE}}}}^{{{\text{off,}}\min }} - 1}} {\left( {I_{{{\text{DE,}}t}} + M_{{{\text{DE,}}t}} } \right)} \le 1$$

(14)

$$\sum\limits_{t = k}^{{k + N_{{{\text{DE}}}}^{{{\text{on,}}\max }} }} {U_{{{\text{DE,}}t}} } \le N_{{{\text{DE}}}}^{{{\text{on,}}\max }} ,I_{{{\text{DE,}}t}} + M_{{{\text{DE,}}t}} \le 1$$

(15)

$$I_{{{\text{DE,}}t}} - M_{{{\text{DE,}}t}} = U_{{{\text{DE,}}t}} - U_{{{\text{DE,}}t - 1}}$$

(16)

$$U_{{{\text{DE,}}t}} P_{{{\text{DE}}}}^{\min } \le P_{{{\text{DE,}}t}} \le U_{{{\text{DE,}}t}} P_{{{\text{DE}}}}^{\max }$$

(17)

$$- R_{{{\text{DE}}}}^{{{\text{down}}}} \Delta t \le P_{{{\text{DE,}}t}} - P_{{{\text{DE,}}t - 1}} \le R_{{{\text{DE}}}}^{{{\text{up}}}} \Delta t$$

(18)

Constraints (13), (14) and (15) limit the minimum ON, minimum OFF, and maximum ON time periods of the generator, respectively. Non,min DE/Non,max DE denotes the minimum/maximum ON time periods of the generator, and Noff,min DE indicates the minimum OFF time periods of the generator. Constraint (16) relates the startup/shutdown state to the ON/OFF state of the generator. Constraints (17) and (18) limit the output and ramping rate of the generator, respectively. Pmax DE/Pmin DE is the maximum/minimum output power of the generator, and Rup DE/Rdown DE represents the limit of the ramp up/down rate of the generator.

Energy storage constraints

$$U_{{{\text{ES,}}t}}^{{{\text{ch}}}} + U_{{{\text{ES,}}t}}^{{{\text{dis}}}} \le 1$$

(19)

$$0 \le P_{{{\text{ES,}}t}}^{{{\text{ch}}}} \le U_{{{\text{ES,}}t}}^{{{\text{ch}}}} P_{{{\text{ES}}}}^{{\text{ch,max}}} ,0 \le P_{{{\text{ES,}}t}}^{{{\text{dis}}}} \le U_{{{\text{ES,}}t}}^{{{\text{dis}}}} P_{{{\text{ES}}}}^{{\text{dis,max}}}$$

(20)

$$S(t) = S(t - 1) + P_{{{\text{ES,}}t}}^{{{\text{ch}}}} \Delta t\eta_{{\text{C}}} - P_{{{\text{ES,}}t}}^{{{\text{dis}}}} \Delta t/\eta_{{\text{D}}}$$

(21)

$$S^{\min } \le S(t) \le S^{\max } ,S(0) = S(N_{{\text{t}}} )$$

(22)

Constraint (19) prohibits the simultaneous charging and discharging of the energy storage at any time period, where Uch ES,t/Udis ES,t is the charging/discharging state of the energy storage at t. Constraint (20) states the maximum charge/discharge power of the energy storage, and (21)–(22) maintain the state of charge (SOC) of the energy storage within the setting range and the cyclic operation of the energy storage. Pch,max ES/Pdis,max ES is the maximum charge/discharge power of the energy storage; S^max/S^min denote the maximum/minimum SOC of the energy storage; S(t) is the SOC of the energy storage at period t; and S(0)/S(N_t) is the initial/final SOC of the energy storage in each scheduling cycle.

Grid-connected line constraints

$$U_{{{\text{GL,}}t}}^{{{\text{buy}}}} + U_{{{\text{GL,}}t}}^{{{\text{sell}}}} \le 1$$

(23)

$$0 \le P_{{{\text{GL,}}t}}^{{{\text{buy}}}} \le U_{{{\text{GL,}}t}}^{{{\text{buy}}}} M_{{{\text{GL}}}}^{{{\text{buy}}}} ,0 \le P_{{{\text{GL,}}t}}^{{{\text{sell}}}} \le U_{{{\text{GL,}}t}}^{{{\text{sell}}}} M_{{{\text{GL}}}}^{{{\text{sell}}}}$$

(24)

Expressions (23)-(24) support the constraints on the grid-connected line’s trading direction and power capacity, respectively. Ubuy GL,t/Usell GL,t is the buying/selling state of the grid-connected line at period t, and Mbuy GL/Msell GL denotes the maximal buying/selling power of the grid-connected line.

Demand response constraints

$$0 \le P_{{{\text{L,}}t}}^{{{\text{cut}}}} \le P_{{{\text{L,}}t}}^{{\text{cut,max}}} ,0 \le P_{{{\text{L}},t}}^{{{\text{tran}}}} \le P_{{\text{L}}}^{{\text{tran,max}}}$$

(25)

$$\sum\limits_{{t = t_{{\text{L}}}^{1} }}^{{t_{{\text{L}}}^{{{\text{end}}}} }} {P_{{{\text{L}},t}}^{{{\text{tran}}}} \Delta t} = E_{{\text{L}}}^{{{\text{tran}}}}$$

(26)

Constraint (25) limits the power of the sheddable and transferable loads, where Pcut,max L,t is the maximum power for load shedding at t; Ptran,max L is the maximum power of transferable load; and Ptran L,t denotes the power of transferable load at period t. There are some transferable loads, such as household appliances, e.g., washing machines. The total power consumptions of these loads are relatively stable, that is, the planned energy of transferable load Etran L is constant, while their operating time can be regulated within the defined intervals [t1 L, tend L] based on the users’own needs. Thus, (26) enforces the total electricity consumption of the transferable load. In real-world cases, the operational characteristics of demand response are much more complex, including factors such as runtime continuity and user preferences. These aspects are not fully captured in constraint (26), and detailed modeling of demand response will be explored in future work based on practical scenarios.

Power balance constraints

$$P_{{{\text{PV,}}t}} + P_{{{\text{DE,}}t}} + P_{{{\text{GL,}}t}}^{{{\text{buy}}}} - P_{{{\text{GL,}}t}}^{{{\text{sell}}}} = l_{t}^{ * } - P_{{{\text{L,}}t}}^{{{\text{cut}}}} + P_{{{\text{L}},t}}^{{{\text{tran}}}} + P_{{{\text{ES,}}t}}^{{{\text{ch}}}} - P_{{{\text{ES,}}t}}^{{{\text{dis}}}}$$

(27)

Constraint (27) guarantees the power balance for the power grid with multiple devices, where l* t is the predicted value of load demand for each period. In contrast, the 30-min real-time dispatch modifies the 24-h-ahead schedule based on short-term prediction data and aims to minimize operating expenses over the subsequent N_h periods. The formulation of the 30-min real-time dispatch model is as follows:

30-min real-time dispatching objective

$$\mathop {\min }\limits_{\substack{ U_{{{\text{ES,}}h}}^{{{\text{ch}}}} ,U_{{{\text{ES,}}h}}^{{{\text{dis}}}} ,U_{{{\text{GL,}}h}}^{{{\text{buy}}}} ,U_{{{\text{GL,}}h}}^{{{\text{sell}}}} , \\ P_{{{\text{PV,}}h}} ,P_{{{\text{DE,}}h}} ,P_{{{\text{ES,}}h}}^{{{\text{ch}}}} ,P_{{{\text{ES,}}h}}^{{{\text{dis}}}} ,P_{{{\text{L,}}h}}^{{{\text{cut}}}} ,P_{{{\text{GL,}}h}}^{{{\text{buy}}}} ,P_{{{\text{GL,}}h}}^{{{\text{sell}}}} } } \left( \begin{gathered} C_{{{\text{DE}}}}^{{{\text{fuel}}}} + C_{{{\text{DE}}}}^{{{\text{O}}\& {\text{M}}}} + C_{{{\text{ES}}}}^{{{\text{O}}\& {\text{M}}}} + C_{{{\text{PV}}}}^{{{\text{O}}\& {\text{M}}}} \\ + C_{{{\text{ES}}}}^{{{\text{loss}}}} + C_{{\text{L}}}^{{{\text{cut}}}} + C_{{{\text{GL}}}}^{{{\text{trade}}}} + C_{{{\text{ES}}}}^{{{\text{pena}}}} \\ \end{gathered} \right)$$

(28)

where:

$$C_{{{\text{DE}}}}^{{{\text{fuel}}}} = m_{{{\text{DE}}}}^{{{\text{fuel}}}} \sum\limits_{h = 1}^{{N_{{\text{h}}} }} {\left( {a_{{{\text{DE}}}} P_{{{\text{DE,}}h}} + b_{{{\text{DE}}}} P_{{{\text{DE}}}}^{{\text{r}}} U_{{{\text{DE,}}t}}^{*} } \right)} \Delta h$$

(29)

$$C_{{{\text{DE}}}}^{{{\text{O}}\& {\text{M}}}} = m_{{{\text{DE}}}}^{{{\text{O}}\& {\text{M}}}} \sum\limits_{h = 1}^{{N_{{\text{h}}} }} {P_{{{\text{DE,}}h}} } \Delta h$$

(30)

$$C_{{{\text{ES}}}}^{{{\text{O}}\& {\text{M}}}} = m_{{{\text{ES}}}}^{{{\text{O}}\& {\text{M}}}} \sum\limits_{h = 1}^{{N_{{\text{h}}} }} {\left( {P_{{{\text{ES,}}h}}^{{{\text{ch}}}} + P_{{{\text{ES,}}h}}^{{{\text{dis}}}} } \right)} \Delta h$$

(31)

$$C_{{{\text{PV}}}}^{{{\text{O}}\& {\text{M}}}} = m_{{{\text{PV}}}}^{{{\text{O}}\& {\text{M}}}} \sum\limits_{h = 1}^{{N_{{\text{h}}} }} {P_{{{\text{PV,}}h}} } \Delta h$$

(32)

$$C_{{{\text{ES}}}}^{{{\text{loss}}}} = m_{{{\text{ES}}}}^{{{\text{loss}}}} \sum\limits_{h = 1}^{{N_{{\text{h}}} }} {\left( {P_{{{\text{ES,}}h}}^{{{\text{ch}}}} + P_{{{\text{ES,}}h}}^{{{\text{dis}}}} } \right)} \Delta h$$

(33)

$$C_{{\text{L}}}^{{{\text{cut}}}} = m_{{\text{L}}}^{{{\text{cut}}}} \sum\limits_{h = 1}^{{N_{{\text{h}}} }} {P_{{{\text{L,}}h}}^{{{\text{cut}}}} } \Delta h$$

(34)

$$C_{{{\text{GL}}}}^{{{\text{trade}}}} = \sum\limits_{t = 1}^{{N_{{\text{h}}} }} {\left( {pr_{{{\text{GL,}}h}}^{ * } P_{{{\text{GL,}}h}}^{{{\text{buy}}}} - \sigma \cdot pr_{{{\text{GL,}}h}}^{ * } P_{{{\text{GL,}}h}}^{{{\text{sell}}}} } \right)} \Delta h$$

(35)

$$C_{{{\text{ES}}}}^{{{\text{pena}}}} = m_{{{\text{ES}}}}^{{{\text{pena}}}} \left| {S(N_{{\text{h}}} ) - S(t)^{*} } \right|$$

(36)

30-min real-time dispatching constraints

PV output constraints

$$P_{{{\text{PV,}}h}} = p_{h}^{ * }$$

(37)

Generat4or constraints

$$U_{{{\text{DE,}}t}}^{*} P_{{{\text{DE}}}}^{\min } \le P_{{{\text{DE,}}h}} \le U_{{{\text{DE,}}t}}^{*} P_{{{\text{DE}}}}^{\max }$$

(38)

$$- R_{{{\text{DE}}}}^{{{\text{down}}}} \Delta h \le P_{{{\text{DE,}}h}} - P_{{{\text{DE,}}h - 1}} \le R_{{{\text{DE}}}}^{{{\text{up}}}} \Delta h$$

(39)

Energy storage constraints

$$U_{{{\text{ES,}}h}}^{{{\text{ch}}}} + U_{{{\text{ES,}}h}}^{{{\text{dis}}}} \le 1$$

(40)

$$0 \le P_{{{\text{ES,}}h}}^{{{\text{ch}}}} \le U_{{{\text{ES,}}h}}^{{{\text{ch}}}} P_{{{\text{ES}}}}^{{\text{ch,max}}} ,0 \le P_{{{\text{ES,}}h}}^{{{\text{dis}}}} \le U_{{{\text{ES,}}h}}^{{{\text{dis}}}} P_{{{\text{ES}}}}^{{\text{dis,max}}}$$

(41)

$$S(h) = S(h - 1) + P_{{{\text{ES,}}h}}^{{{\text{ch}}}} \Delta h\eta_{{\text{C}}} - P_{{{\text{ES,}}h}}^{{{\text{dis}}}} \Delta h{/}\eta_{{\text{D}}}$$

(42)

$$S^{\min } \le S(h) \le S^{\max } ,S(0) = S(t)^{*}$$

(43)

Grid-connected line constraints

$$U_{{{\text{GL,}}h}}^{{{\text{buy}}}} + U_{{{\text{GL,}}h}}^{{{\text{sell}}}} \le 1$$

(44)

$$0 \le P_{{{\text{GL,}}h}}^{{{\text{buy}}}} \le U_{{{\text{GL,}}h}}^{{{\text{buy}}}} M_{{{\text{GL}}}}^{{{\text{buy}}}} ,0 \le P_{{{\text{GL,}}h}}^{{{\text{sell}}}} \le U_{{{\text{GL,}}h}}^{{{\text{sell}}}} M_{{{\text{GL}}}}^{{{\text{sell}}}}$$

(45)

Demand response constraints

$$0 \le P_{{{\text{L,}}h}}^{{{\text{cut}}}} \le P_{{{\text{L,}}h}}^{{\text{cut,max}}}$$

(46)

Power balance constraints

$$P_{{{\text{PV,}}h}} + P_{{{\text{DE,}}h}} + P_{{{\text{GL,}}h}}^{{{\text{buy}}}} - P_{{{\text{GL,}}h}}^{{{\text{sell}}}} = l_{h}^{ * } - P_{{{\text{L,}}h}}^{{{\text{cut}}}} + P_{{{\text{ES,}}h}}^{{{\text{ch}}}} - P_{{{\text{ES,}}h}}^{{{\text{dis}}}}$$

(47)

The meanings of Eqs. (28)-(35) and constraints (37)-(47) are like those in the 24-h ahead scheduling model. The 30-min real-time dispatch model is formulated by replacing t and N_t in the 24-h ahead scheduling model with h and N_h. However, it should be noted that (43) and (36) limit the initial and final SOCs of the energy storage in the 30-min real-time dispatch. The penalty cost of the deviations between the final SOC and the 24-h ahead optimal SOC is added to the objective as (36), which links the 30-min real-time dispatch with the 24-h ahead scheduling. The detailed description of methodology is discussed in the reference³⁶.

Results

Digital solutions during transition

In pursuit of energy transition, the EMA initiated digital initiatives in conjunction with industry partners to ensure the long-term viability of the country’s energy grid infrastructure. The Grid Digital Twin and DERMS have undergone significant advancements, and their progress will persist in the coming years. These proposed solutions contribute to the SG electricity grid’s increased resilience and dependability, while also facilitating the implementation of RE resources. Two models of the Grid Digital Twin comprise: Digital Asset Twin and Digital Network Twin. The former symbolizes the Singapore Power (SP) electricity network assets in the form of a virtual replica of the Digital Twin, while the latter represents sophisticated modeling and simulation software utilized to simulate the grid’s response to EV demand. The objective of the network is to create an EMA-compatible end-user software solution. By 2025, the complete counterpart is intended to be dispatched as a pilot. The digital twin technologies enable the examination and testing of the SG energy grid under a variety of scenarios in a risk-free environment. On the flipside, DERMS is focusing on solar PV forecast and mitigating its intermittency via ESS currently and through EV in the later stage while effectively managing reliability and system costs³⁷.

In view of the above discussion, several possible solutions for the future SG power grid need to be developed which pave the way for net-zero emission by 2050. There is no one solution for the decarbonization of the energy sector in any country including SG. One of the digital solutions called blockchain-based P2P energy trading framework for future SG has been developed by a group of researchers at Clean Energy Research Lab (CERL)³⁸. As SG is new to decentralized energy markets, implementing it will not be a straightforward process and it involves several steps, including regulatory, technological, and market development. This is a complex and multifaceted process that requires collaboration among government bodies, utilities, technology providers, and consumers. It should be tailored to the specific needs and conditions of SG. A pilot project was launched in 2019 by a company, Electrify, in collaboration with Senoko Energy, and is sponsored by Engie Factory, the venture of a French multinational electric utility company. After running the pilot project until June 2021, the company has a near-term goal of deploying the P2P energy trading business models for residential and business clients. Initially, targeting 100 participants including consumers who wish to power their households and trade the RE of the solar PV system installed in their rooftops. In the long term, Electrify said it aimed to set up a grid-wide P2P energy trading technology named Synergy. Because of this, the company had initiated its talk with established energy traders and regional utilities to bring Synergy into the markets³⁹. Similarly, a pilot project has been run successfully for eight months in two phases in Malaysia on P2P energy trading by the Sustainable Energy Development Authority (SEDA) in collaboration with Australian Power Ledger Ltd Pty^40,41. The project provides insights into consumer preferences and price model settings. In a broader sense, it also plays a bigger role in supporting climate change. The pilot run is made to capitalize on its rooftop solar PV potential which is about 37.4 GW in peninsular Malaysia alone⁴². The feasibility of P2P energy markets for various grid services has been tested by the researchers at CERL, NTU for P2P trading within the micro-grid, as well as for voltage and frequency ancillary services.

Rise of sustainable prosumers in future SG

Within the context of SG’s growing sustainable landscape, a noteworthy transformation is unfolding as traditional consumers evolve into proactive prosumers, capable of both consumption as well as production. The prosumer’s equipment encompasses cutting-edge elements. Rooftop solar panels serve as decentralized power generators, producing clean electricity for personal consumption and facilitating the sharing of surplus energy within the community. Simultaneously, EVs support sustainability by engaging in V2G activities during peak demand periods, stabilizing grid requirements, and contributing surplus power back to the power grid. Integration of Internet-of-Things (IoT) facilities and smart meters provides real-time insights into energy consumption and production, enabling informed energy management and further optimizing operational costs.

Empowering prosumers with advanced technology, the P2P energy marketplace facilitates efficient energy sharing and cost savings. Prosumers directly trade surplus solar power with neighbors, creating a micro-economy that boosts local energy resilience and reduces reliance on distant power sources. Additionally, they optimize energy consumption by procuring electricity from neighbors at potentially lower rates, promoting cost efficiency and responsible energy use. The dynamic P2P market with real-time pricing contributes to grid stability, fostering a sense of community and shared responsibility for sustainable practices. In SG, the prosumer paradigm aligns seamlessly with ambitious environmental goals. Initiatives like the Green Plan 2030, led by the EMA, create an environment for flourishing P2P markets. The envisioned interconnected neighborhoods contribute to a sustainable energy ecosystem, with the prosumer acting as a key driver for cleaner, more resilient, and equitable energy landscapes in megacities like SG.

SIEW Energy Insights highlights the potential of P2P energy trading as a viable alternative to enhance accessibility to RE in SG. A collaborative effort led by Senoko Energy, ENGIE, and Electrify resulted in the successful launch of SolarShare, a P2P solar energy trading platform that matches real-time solar energy generation with usage. During the trial, SolarShare facilitated consumers in sourcing 43% of their electricity from solar power, showcasing its effectiveness. The platform, supported by digitalization and smart meters, enables users to set and match preferred energy rates seamlessly. It has also been emphasized that the platform’s potential for further improvement through the integration of blockchain and ESS. This innovative approach marks a significant step towards realizing SG’s clean energy ambitions.

End user application for future energy transition

With the rising concern of RE based DERs integration into the existing electrical infrastructures and the need to trade the power within the neighbors in the community, a P2P trading based mobile application has been developed with an initiative of trading power by the utility consumers at the CERL. P2P energy trading is considered one of the potential solutions for a city-state like SG to optimize the use of DERs, reduce stress on the grid, and offer incentives to build rooftop solar photovoltaics. The reliability of the developed Android app is verified by testing it for P2P energy trading in the 0.4 kV micro-grid setup at CERL. A Flutter-based Android app that uses Dart as its programming language for P2P energy trading is built on the blockchain system, and it fully utilizes information about the participating prosumers and consumers, along with their electricity prices and quantities. The app supports energy trading by storing transaction information in blocks, verifying transaction validity by all nodes in the network, and ensuring transaction security and privacy through encryption. A prototype mobile application was developed to demonstrate the P2P energy trading process. The app runs on Android. To validate the proof of concept, a lab-scale microgrid testbed was used in CERL at NTU. In this setup, the microgrid generation sources acted as prosumers and controllable loads as consumers. The market server was deployed within a local area network (LAN), with each asset (generation/load) assigned a static IP for secured communication. Transactions between peers were executed and recorded via the blockchain ledger to ensure data integrity and transparency. This setup serves as a prototype demonstration, not a full market simulation, but it illustrates the technical feasibility and scalability of the approach. We believe the proposed architecture has strong potential for real-world implementation, especially in decentralized renewable energy systems involving small to medium-scale producers such as PV and ESS operators.

Figure 5(a) shows the login page for the developed P2P energy trading app in an Android mobile system. On this page, we have designed the application with a simple design such as the NTU logo, the name of the application, as well as the fields to fill in the user’s email and password. Users who wish to use this app for trading can download and register using their email addresses to secure access. Later, users could enter the market and access various information on the current trade, assets that they have, and the transaction history with the buttons via navigation. For instance, the app has been tested with the micro-grid setup at CERL considering the prosumers and consumers participating in the trade, and the results are depicted in Fig. 5(b). The screenshot represents all the transactions done via the energy market using the unified units of energy. They serve as general units to which the user can refer to. Every bid submitted by the account using the mobile application will be reflected on the blockchain at the backend. If the trading is successful, the transaction records will be reflected in the trade history page of the mobile application. The blockchain backend interface is shown in Fig. 5(c). When the energy transaction is executed successfully, the interface will demonstrate the hash code of the new block as well as the account addresses of buyers, sellers, and smart contracts. The entire operation has been recorded and uploaded, and an open-source NTU P2P Trading (CERL) mobile app can be downloaded from the Google Play Store⁴³. They may attract investors to further enhance the application of P2P trading.

Fig. 5

Blockchain-based P2P energy trading mobile application for end users.

In future, the prosumers participating in P2P energy trading can also help in supporting the frequency and voltage ancillary services to ensure the reliability of the grid. Also, the mobile app can help in charging/discharging EVs cost-effectively by supporting demand response management. One of the plausible solutions of P2P energy trading for secondary frequency services has been studied by the researchers at CERL³⁵.

The P2P energy trading mechanism, implemented using a Proof-of-Authority (PoA) based private blockchain, also demonstrates scalability and broader applicability. While this study uses a double auction mechanism for market clearing, alternative strategies, such as game-theoretic pricing models, could be incorporated to further optimize market outcomes. The mobile application developed demonstrates local feasibility, but the same architecture can be adapted for decentralized energy markets in other regions, depending on local regulations and market structures.

Grid evolution during transition

Over the years, SG has extensively invested in developing smart grid technologies and ensuring that its power infrastructures are among the most stable globally. SG has initiated the development of cost-competitive energy solutions to increase energy efficiency, decrease carbon emissions, and diversify energy sources. The energy transition will necessitate a shift to Smart Grid 2.0, a new generation of electricity distribution that enables two-way flows of energy and communications and transforms the electric power grid into an interoperable network by integrating information and communication technologies with the power-delivery infrastructure⁴⁴. Given this, JTC corporation and SP group have developed the first smart grid for SG implementation in Punggol Digital District (PDD). It leads to a reduction of carbon emissions by 1,700 tonnes annually which is equivalent to taking 270 cars off the road. Further, EMA in collaboration with the Singapore Institute of Technology (SIT) is trying to develop innovative solutions to manage rising DERs interconnection to ensure grid reliability and stability. The smart grid will be integrated with an open digital platform (ODP) allowing interconnection among the buildings in the district. At the forefront of the SG smart nation push, PDD is intended to be a vibrant and comprehensive district where trailblazing technology and social innovation transform will meet the smart nation ambition of future SG. This helps to develop the digital twin of the grid using the ODP to simulate the scenarios for future development⁴⁵.

Intermittency mitigation during transition

SG is in tropical regions and undergoes variations in climatic conditions due to the tropical rainforest climate, which means it is generally hot and humid throughout the year with regular rainfall. Hence, the major installed renewable energy sources (RES) of PV farms undergo large fluctuations in the output power from the panel. Earlier, the SG government announced the plans for solar PV power with a capacity of 1 GWp beyond 2020 to enhance the energy mix. However, based on an EMA published report, a solar capacity of over 820 MWp has been reached by the end of 2022. Later, in March 2020 the EMA reviewed and announced a quadratic growth of solar in SG to 2 GWp by 2030 and a linear growth to 4 GWp by 2050. However, the target by 2025 is to have a capacity of 1.5 GWp, recognizing the benefits of solar generation rather than importing the power from other countries. The maximum electricity demand is forecasted to be 9,300 MW in 2030 meeting the electricity demand of around 350,000 households⁴⁶. To facilitate the entry of more solar PV prosumers to participate into the electricity market, the government has introduced the Enhanced Central Intermediary Scheme (ECIS) to sell the excess power to the grid. This scheme allows the consumers to sell the excess generation of Intermittent Generation Sources (IGS) or the non-IGS below 10 MW to sell excess power through an intermediary (SP services) instead of the Energy Market Company (EMC)⁴⁷. Additionally, SG has initiated the application of AI to grid operations to predict short-term supply and demand. At present, the AI system has acquired knowledge from historical weather data, real-time solar irradiance measurements, satellite images of cloud cover, and historical weather information to predict solar output and devise generation and storage dispatch schedules. Additionally, electricity demand is predicted by the system using historical consumption data. By the middle of the 2030 s, both combinations of the forecast will assist SG in reducing its dispatch window from 30 to 5 min.

The Intermittency Pricing Mechanism (IPM) white paper was published by the EMA in October 2017. This article details the IPM, which is used to fairly charge all forms of generation for their share of reserves. The IGS includes RESs like solar photovoltaic and wind energy, where the output power oscillates due to uncertainties in solar irradiance and wind profiles. Due to space constraints and a low average wind speed of 0.8 m/s, the wind energy system is unsuitable for SG. However, the IGS, like solar PV, has a larger scope in SG and the merits of reduced carbon emission with zero fuel cost can contribute towards the overall reduction of carbon emissions in SG. The carbon tax has been recently revised, and its value keeps rising as given in Fig. 6. The revised carbon tax is critical in empowering the pace of transformation to achieve the climate ambition of net zero emission. This will also help the business in developing the solution which ensures the new investments, policies and economic activities that are in alignment with the national climate goals and are competitive in a low-carbon future. Similar carbon pricing schemes are implemented in several countries like EU, South Korea, and California, USA⁴⁸.

Fig. 6

SG Carbon tax trajectory⁴⁸.

In addition, to ensure the growth of IGS, an appropriate pricing mechanism to allocate the reserves in a fair manner is introduced via Intermittent Pricing Mechanism (IPM) by EMA. In general, the information of the magnitude of power loss and frequency is considered for every half-hourly period to distribute the reserve costs. The IPM is implemented by integrating all IGS and the cost for mitigation is shared with the individual IGS based on their installed capacity. In SG, the IPM is preferably applicable to the non-residential consumers than the residential consumers⁴⁹.

Since the power output of IGS fluctuates more randomly, even in the case of aggregated prosumers, the power system still poses a severe burden to balance the output of IGS. Standby generators and energy storage devices are necessary to mitigate the intermittencies posed by PV plants due to the frequent cloudy and rainy weather conditions which may cause grid reliability issues. To improve grid resilience, the government has initiated the deployment of a 285-MWh ESS capacity on Jurong Island. This will ensure the security and reliability of power output to the consumers. This encourages businesses to spend money on solutions like energy storage and demand-side management that can deal with intermittent power supplies. Within a thirty-minute dispatch window, if the IGS maximum output decreases, then the predicted maximum power loss can be used to determine the reserve cost. Researchers in power engineering can see that the reserve cost can be reduced if solar PV implements intermittency mitigation methods to minimize the predicted maximum power reduction. In view of this, a study has been made by the research team at CERL in NTU to develop a hierarchical energy management system (Hi-EMS) which can help to overcome the intermittency by appropriately scheduling of reserves based on the PV output power. This contributes to the EMA digital solutions of DERMS project for solar PV forecast and intermittency mitigation.

The Hi-EMS module aims to develop a decision-making tool that will produce control signal commands for distributed generation (DG) and battery energy storage system (BESS) optimization. The objective is to reduce operation expenses and mitigate solar intermittency while achieving coordination among all power units. The degradation cost of BESS, the start-up, and shut-down costs of DGs, the operation and maintenance costs of PV, and the trading cost with the main grid are all included in the total operation cost. To mitigate the effects of anticipated fluctuations in photovoltaic (PV) outputs, a framework for bilevel rolling optimization is suggested. Rolling optimization functions on the following two-time scales:

24-hour ahead scheduling: With an interval of Δt = 30 minutes, the renewable and load power of the N_t time periods (N_t = 48) over the subsequent twenty-four hours are predicted. The operational constraints of the system and equipment, as well as the 24-hour-ahead prediction data, are utilized to optimize the power plans of different units (e.g., the DG, BESS, and grid-connected line (GL) representing power import/export). The objective is to minimize the overall cost of operation and scheduling. The referral plans for the 30-minute real-time dispatch are derived from the optimal results of the time coupling plans, which are obtained once every day.

Real-time 30-minute dispatch: Predictions of the renewable and load power for the subsequent N_h periods (N_h = 6) are made using the time interval Δh = 5 minutes. Leveraging real-time prediction data and 24-hour-ahead referral plans, the operation/dispatch plans for the subsequent six periods are optimized to reduce dispatch costs. In each real-time optimization, the initial state of charge (SOC) of the BESS is determined by the optimal final SOC of the real-time dispatch during the previous period. When the objective cost function connects the two-time scales, a penalty cost is applied to account for the discrepancies between the final SOC in each real-time dispatch and the referral SOC in the 24-hour advance schedule. The real-time optimization is executed at 5-minute intervals, with the dispatch instructions being determined by the optimal outcomes of the initial period.

The Hi-EMS developed in this study demonstrates a high degree of generalizability and can be adapted to various application scenarios. Firstly, in terms of system architecture, the combination of 24-h ahead scheduling and 30-min real-time dispatch effectively captures the intermittency of renewable energy generation across different timescales. Given that electricity markets in different regions operate at varying time resolutions, the optimization timescale of the Hi-EMS can be readily adjusted to suit specific market conditions. Secondly, this study focuses on the coordination of distributed generation (DG), energy storage systems (ESS), and the main grid to smooth the fluctuations of variable renewable energy. If other flexible power generation units exist in the system, their operational models can be integrated into the Hi-EMS based on their scheduling characteristics without altering the fundamental structure of the framework.

Regarding ESS strategies, the Hi-EMS incorporates battery degradation costs and formulates scheduling decisions with the objective of minimizing the overall system operation cost. In practical implementations, different strategies may be adopted, for instance, determining charging and discharging profiles based on net load, dynamic pricing, or ancillary service requirements. The Hi-EMS framework is flexible enough and can accommodate such variations by modifying the relevant constraints and objective functions accordingly.

A typical scenario is tested by assuming that the PV generation is relatively sufficient at noon compared to the total load demand. Figure 7(a) and (b) plot the predicted electricity prices, PV power, and load demand on one typical day in September 2022. The peak load power within the day is normalized to 1 p.u.. Figure 7(c) depicts the BESS and DG power dispatch commands from Hi-EMS to meet the load demand to mitigate the PV intermittencies. Further, a shiftable load as a demand response (DR) participant is incorporated in this case. The total energy requirement of this shiftable load is assumed to be 0.24 p.u., and the allowable shifting time is 6 pm to 11 pm. Also, the maximal power of the shiftable load is 0.12 p.u.. As seen in Fig. 7, the PV has no output power at night, hence there are no significant power fluctuations. The BESS (also called Battery) and DG have almost shut down due to their high operation costs compared to the electricity prices. During the daytime, the DG, BESS, and PV are coordinated for load supply to obtain the economic dispatching plans, and the long-term net load fluctuations caused by PV are compensated using the BESS and DG to reduce their impact on the main grid.

Fig. 7

Effective PV power output on a sunny day to meet the demand.

Figure 7(d) gives the optimal load power plans with and without the DR. Compared to the operation cost without the DR, the DR results in a 5.7% increase in the overall profit. Moreover, the optimal power results of the grid-connected line (GL) power with and without intermittency mitigation are given in Fig. 7(e). Without mitigation, the power fluctuations of the net load are consistent with the PV output. By making full use of the BESS and DG, the fluctuations of the net load are greatly reduced, and the power of the GL during the daytime is stable after mitigation. The optimal cost objective with and without mitigations shows a 7.2% reduction in the overall payment.

Another case study is conducted by assuming that the PV generation is always below the load demand in the daytime. Figure 8(a) and (b) illustrate the predicted electricity prices, PV power, and load demand on one typical day in September 2022. Figure 8(c) depicts the BESS and DG power dispatch commands from Hi-EMS to meet the load demand to mitigate the PV intermittencies. Further, we consider the shiftable load as DR participants between 8 am to 1 pm. The total energy requirement is 0.16 p.u., and the maximal power of the shiftable load is 0.08 p.u. According to Fig. 8(d), the shiftable load is transferred to 9 am-11 am as the PV generation is insufficient for the load supply. The shiftable load is transferred to these periods with lower energy prices. This reduces the total operation cost of the dispatching plans. Moreover, the shiftable load also helps to smooth the power fluctuations from 9 am-11 am caused by PV. The injected power into the main grid, GL power, is stable after mitigation, as can be seen in Fig. 8(e). The results show a 5.8% increase in the overall profit due to the adoption of DR and a 6.9% reduction in the overall cost for the PV intermittency payment.

Fig. 8

PV power output on a cloudy day ineffective to meet the demand.

To further validate the effectiveness of the developed Hi-EMS, especially in response to transitory shocks from PV outputs, we conduct optimization tests on the other four typical days. Figure 9 presents the optimal dispatch commands from Hi-EMS for the four typical days. As shown in Fig. 9(b), the PV power exhibits obvious intermittency between 1 and 4 pm. To ensure the reliable load supply, excess solar generation is injected into the main grid through GL, and the DG is started up during the periods of insufficient PV generation. It can be seen from Fig. 9(c), PV power experiences a sudden drop from 1 p.u. to 0.5 p.u. at 11 am due to the weather conditions. To mitigate the transitory shock, the power of DG generation compensates for the power deficit. Similarly, in Fig. 9(d), the PV output rapidly decreases at 12 pm, and the system uses the DG generation and the purchased power from the main grid to meet the power shortage. For the four typical scenarios, our results respectively indicate an 8.1%, 9.8%, 4.2%, and 6.3% reduction in the overall cost for the PV intermittency payment.

Fig. 9

Optimal dispatch commands from Hi-EMS for the other four typical days.

It is quite evident that the developed Hi-EMS presented from the above case studies can reduce the IPM payment. The solution is a part of energy digitalization towards the DERMS project which mitigates the PV intermittencies for reliability of the supply. Advancing energy management to a further level by considering the EV scenarios may improve the mitigation costs and benefit the consumers in general to use ample amounts of RE sources. In addition to providing digital solutions to improve the grid infrastructure, SG made other initiatives in deploying the ESS for future power reserve and trying to find an alternative source of hydrogen power generation within the nation.

Storage deployment during transition

One of the prominent RESs in SG is solar PV power. Due to the rain and cloud covers in the tropical region of SG, it is difficult to harvest power even during the daytime. To mitigate the intermittent issues in solar power, one feasible solution is to install the ESS to maintain grid reliability. Several industries have collaborated with the EMA on ESS deployments and developed R&D testbeds. A partnership between EMA and the SP group in October 2020 will deploy SG’s first utility-scale ESS with a capacity of 2.4 MWh at a substation as an initiative. It can satisfy the daily energy needs of 200 Housing & Development Board (HDB) four-room households. It is expected that the distributed ESS installed in the electrical switch rooms of HDB will prove to be a viable solution for SG in the future, enabling greater integration of solar energy into the grid. In light of this, EMA, in conjunction with Sunseap (Now EDPR APAC), intends to implement ESS in five HDB blocks in Punggol as demonstration sites to address the issue of solar PV intermittency. Future deployments will be better equipped to participate in the wholesale electricity market and offer services to mitigate intermittency and decrease peak load demand thanks to the knowledge acquired during this initiative⁵⁰. On the other hand, a 285 MWh of ESS has been installed in Jurong Island by EMA in collaboration with Sembcorp and its operation started in December 2022. This achieves the SG’s target of reaching at least 200 MWh ESS beyond 2025 ahead of time. It is identified as one of the largest ESS in Southeast Asia, and among the fastest of its size to be deployed⁵¹. The Floating Living Lab has been developed by offshore and marine company, Seatrium, at its Pioneer Yard and is the first floating and stacking lithium-ion based ESS in the nation. It will begin its operation in the first quarter of 2024 with the maximum capacity of 7.5 MWh which can meet the need of 600 four-room HDB flats in a single-discharge. To enhance its operation the ESS is integrated with the smart energy management system that uses AI and ML to enhance its efficiency. The findings from the testbed help to deploy ESS in SG shortly to ensure the resilience and stability of the grid⁵². In addition, an ice thermal ESS will be installed at the George Street substation by EMA and SP group as a pilot project to optimize the space and it will commence its operation in the third quarter of 2026. This will enable the curtailment of 2-MW load. SP can potentially enhance its capacity curtailment capabilities for DR during peak hours and intermittency mitigation by augmenting the quantity of storage containers⁵³. To minimize overall project costs due to scarcity of land in SG, Fig. 10 represents the possible deployment of ESS in SG⁵⁴.

Fig. 10

ESS deployment plans in future for SG energy transition⁵⁴.

Hydrogen power development during transition

SG is working closely with industry and international partners, to realize the low-carbon hydrogen technology to the island nation. It is expected that 50% of SG demand will be supplied by hydrogen power by 2050⁵⁵. SG is anticipated to get its first hydrogen-ready power plant in Jurong Island by the first half of 2026. The Keppel Sakra Cogen Plant in Jurong Island will be running on clean hydrogen in the future and can supply a power of 600 MW of electricity which amounts to 9% of the peak electricity demand of 2020. Currently, the plant is running using natural gas and is designed to operate for fuel with 30% hydrogen content, which produces fewer fossil fuel emissions, and with hydrogen as a fuel in the future without blending. Further, an agreement with Mitsubishi Heavy Industries has been made to study the feasibility of developing a 100% ammonia-fueled power plant that burns CO₂-free like hydrogen. Together, it helps the island nation as well as Southeast Asia to build a more resilient and sustainable energy sector. Upon completion of the project successfully, Keppel’s power generation portfolio will rise from the existing 1300 MW to 1900 MW contributing to the rising need for reliable energy with SG economic development^56,57. Further, an initiative has been made to transport low-carbon hydrogen fuel between the Malaysian state Johor and City Energy’s Senoko gasworks plant in SG via a pipeline. But, the import may be blue hydrogen or green hydrogen as it becomes operational by 2027⁵⁸.

Reduced GHG emission during transition

The reduction of GHG emission for various possible scenarios of energy mix by 2050 are illustrated in Fig. 11. The plots depict a rapid decline in the GHG emission in power sector by 2050 compared to the level of 0.417 kg CO₂/kWh in 2022 on a national level with all three possible scenarios that have been discussed earlier. This reduction in the emission is in line with the Paris agreement of limiting the temperature to 1.5 °C by 2050, with zero emissions in power sectors of SG. The energy sector drives the energy transition in other sectors and SG is taking the initiatives to reduce the GHG emissions in the aviation and transportation industry as well. In particular, the increased impetus on EVs will be a fundamental enabler of climate-compliant energy pathways in SG. The reduction in GHG emissions in the island nation helps to maintain the temperature of the country so that the rise in the sea level in future years can be reduced. This results in social and economic benefits for SG.

Fig. 11

GHG emission and grid emission factor for various possible scenarios of 2050.

Challenges during transition

In this research, we have discussed the various possible scenarios for the transition of the SG power sector. Within the scope of the study, the available renewable sources cannot meet the growing energy demand and henceforth power would need to be imported from the neighboring countries to meet the energy demand and satisfy the target ambition of zero emissions by 2050. Despite space limitations in installing the PV panels, the modeling of future power systems poses uncertainties and challenges in integrating inertia-less renewables while importing power hinges on the government policies made.

• The island nation is hindered by alternative energy sources, land scarcity, and concentrated population faces significant obstacles during the energy transition. SG is, nevertheless, exerting considerable effort to resolve its energy trilemma to provide Singaporeans with energy that is secure, affordable, and sustainable. The Ministry of Trade and Industry recently announced that importing energy from the neighboring countries by an amount of 6 GW by 2035 instead of 4 GW initially, to meet the local energy demand. In view of this, a Conditional Licenses have been issued to import power of 2 GW from Indonesia, and a Conditional Approvals have been issued to import power of 3.6 GW of low-carbon electricity from Cambodia, Indonesia and Vietnam. Further, EMA has issued a Conditional Approval to Sun Cable to import 1.75 GW of energy from Australian Northern Territory to Singapore. In the meantime, cross-border electricity trading is doubled the capacity of electricity traded under the Lao PDR-Thailand-Malaysia-Singapore Power Integration Project to 200 MW. As interconnectors are critical in the path of electricity importing journey, hence, SingaporePower PowerInterconnect (SPPI) is developed as a subsidiary of SP Group to provide collaboration with importers to implement the interconnector projects. Progressively increasing the power production locally from a natural gas, and by implementing solar PV installation at the rooftop and reservoir with capacity of about 2GWp by 2030 and have the current installed capacity of 1.35 GWp and using some low-carbon alternatives⁵⁹.

• Beyond ASEAN, the city-state is strengthening its partnership with international stakeholders. Regarding this, Singapore have officially launched the International Energy Agency (IEA)’s Regional Cooperation Centre, which is the first office located outside of its headquarters in Paris. It is expected that it will be helpful to accelerate regional decarbonization and strengthen the multilateral power trade. Further, Singapore in collaboration with US forming a working group comprising of regional countries and multinational development institutions to advance the study on Legal and Financing Frameworks for Regional Energy Connectivity in Southeast Asia^59,60.

• Despite generation possibilities, companies also have an important role in reducing the carbon emission from electricity consumption. So, the government implemented the Renewable Energy Certificates (RECs) to purchase for the companies operating in city-state to show that the consumption of 1MWh power is from renewable energy sources. Within Singapore, there are standards that protect the credibility of RECs issued from renewable energy projects. However, there is no international policies or standards that recognize RECs associated with cross-border energy trading. Hence, Singapore started working with the Asia Clean Energy Coalition (ACEC), Climate Group’s RE100 initiative, SEMI and the I-TRACK Standard Foundation to co-create framework to recognize RECs associated with cross-border energy trade. In view of this, the EMA in collaboration with EnterpriseSG and the Energy Research Institute @ NTU developing a Playbook for business in Singapore⁶⁰.

• Construct a multi-layered grid that optimizes its operations and reliability through the implementation of digital solutions that enhance grid planning and functioning. This will facilitate enhanced coordination and cooperation among the grid, demand, and supply. Incorporating plug-and-play functionalities to accommodate emerging standards and technologies can potentially enhance the flexibility of pre-investments in infrastructure.

Importing energy may be an easy solution for SG but the cost is high compared to power generation in SG. This also depends on the government policies and energy import leads to many uncertainties. For instance, if a country or country decides to ban clean energy exports, then the demand becomes unsatisfied for a city-state which highly depends on imports. This also results in energy insecurity, and hence we should decide how to overcome it if such a situation arises in the future with its backup power supply and advanced energy technologies like hydrogen and carbon storage⁶¹.

Applicability of energy transition to other city-states

Despite Singapore’s unique aspects, several elements of Singapore energy transition have broader applicability with varying political and economic contexts:

Maximizing urban space for renewable energy

Like SG, many other densely populated city-states, such as Hong Kong, Tokyo, and New York, face similar land constraints. However, the import of energy from neighboring regions to meet the local demand depends on the government policies of individual city-state. On the other hand, the approach of leveraging urban structures through solar PV installation in rooftops and reservoirs can be adopted to other city-states.

Reliance on regional interconnections for energy security

: Due to the limited space constraints, SG government had agreements with ASEAN neighbors in early-stage for cross-border energy trade to meet the load demand, offering insights for regions aiming to enhance energy security while transitioning to renewables (Welsch et al., 2018)⁶². This helps in energy transition and reduces the dependency on domestic fossil fuels. Similar projects of energy interconnection with neighboring region helps another city-state around the world as well to achieve net-zero emission.

Policy and incentive frameworks for private sector involvement

SG is the first country to implement carbon pricing in Southeast Asia and ASEAN to support green energy transition. Also, taken an initiative to move towards Electric Vehicle (EV) by deploying 60,000 EV chargers by 2030. Since the electricity market is centralized, in future energy policy may be introduced to use EV and the used batteries for grid ancillary services like frequency regulation, contingency reserve and demand response via a regulatory sandbox as announced by EMA in the SIEW, 2024. Furthermore, implementing green certification programs, incentive structures for public–private partnerships and investment in RE technologies can serve as a model for other global cities to achieve energy transition. These smart grid initiatives offer a framework adaptable to varying levels of renewable penetration in diverse regions.

Grid emission factor

The amount of carbon emission per unit of electricity generated including all the generation sources like natural gas, solar, and waste-to-energy⁶³. The SG demand has increased from 42 TWh in 2009 to 53 TWh in 2020 at a compound annual growth rate (CAGR) of 2.2%, and its value is 1.8% for the growth of peak demand over the same period. In 2020, the demand fell by 2% due to COVID-19, as the business activities have been dropping rapidly. However, in 2021 the demand is expected to increase by 3% from 2020 and the system peak demand is around 7667 MW. The CAGR has been calculated during this period keeping the demand in 2009 as a base value. For the period of 2022 to 2032, the load demand of year 2022 is considered as the base electricity consumption of 54.9 TWh⁶⁴. In view of this, the annual system demand and peak demand are projected to grow at a CAGR of 2.8–3.2%. The increase in electricity demand is due to the growth of data center and transportation (EV) after COVID-19. With this assumption, the demand has been computed for the upcoming years until 2050 by assuming CAGR to be 3%. During this case, the total demand of SG in 2025, 2030, 2040, and 2050 are 59.99 TWh, 69.55 TWh, 93.52 TWh, and 125.79 TWh, respectively. SG has proposed three possible solutions of generation mix to reach net-zero emission. Table 1 represents the percentage of energy mix for various possible scenarios of SG power energy transition to reach net-zero carbon emission^63,64.

Table 1 Percentage of Energy mix for future SG Energy Transition from 2025 to 2050.

Conclusion

In conclusion, this paper explores the potential integration of both domestic and imported renewable energy sources, particularly solar photovoltaic and energy storage systems, in advancing towards achieving a sustainable and resilient energy future by 2050. This integration helps to reduce the carbon footprint in the city-state via energy mix by enhancing the grid flexibility and resilience to be paramount. Furthermore, cross-border energy trade with ASEAN countries presents both challenges and opportunities; a harmonized regulatory framework and incentivized public–private partnerships will be essential for realizing the vision of city-state as a potential green energy hub in the ASEAN region. Advanced grid management technologies like Hi-EMS are proposed in addition to the P2P energy trading for effective harnessing of renewable resources contributing to the energy transition goals and supporting regional decarbonization efforts. Moving forward, continued research and policy alignment will be essential to overcoming existing technical and regulatory barriers, fostering an ecosystem that supports sustainable energy growth. Through these initiatives, Singapore can emerge as a regional leader in renewable energy integration, setting a benchmark for urbanized nations worldwide.