Smart Microgrid Management and Optimization: A Systematic Review Towards the Proposal of Smart Management Models

Abstract

The increasing integration of renewable energy sources (RES) in power systems presents challenges related to variability, stability, and efficiency, particularly in smart microgrids. This systematic review, following the PRISMA 2020 methodology, analyzed 66 studies focused on advanced energy storage systems, intelligent control strategies, and optimization techniques. Hybrid storage solutions combining battery systems, hydrogen technologies, and pumped hydro storage were identified as effective approaches to mitigate RES intermittency and balance short- and long-term energy demands. The transition from centralized to distributed control architectures, supported by predictive analytics, digital twins, and AI-based forecasting, has improved operational planning and system monitoring. However, challenges remain regarding interoperability, data privacy, cybersecurity, and the limited availability of high-quality data for AI model training. Economic analyses show that while initial investments are high, long-term operational savings and improved resilience justify the adoption of advanced microgrid solutions when supported by appropriate policies and financial mechanisms. Future research should address the standardization of communication protocols, development of explainable AI models, and creation of sustainable business models to enhance resilience, efficiency, and scalability. These efforts are necessary to accelerate the deployment of decentralized, low-carbon energy systems capable of meeting future energy demands under increasingly complex operational conditions.

1. Introduction

The global energy sector is undergoing profound changes driven by the urgent need to achieve carbon neutrality, improve energy security, and support sustainable development. In this context, smart microgrids have become a foundational element for future power systems, enabling the efficient integration of distributed energy resources (DERs) and renewable energy sources (RES) while strengthening system resilience and operational flexibility [1,2]. These localized energy systems support optimized management through advanced control methodologies and real-time decision-making platforms. The growing penetration of variable RES, particularly solar and wind, has increased the challenges related to intermittency and uncertainty, especially in isolated and weakly interconnected grids [3]. Recent developments emphasize the deployment of hybrid energy storage systems, combining battery energy storage systems (BESSs) with hydrogen-based solutions, to meet both short- and long-term storage needs, contributing to improved grid stability and economic performance [4,5].

In parallel, the application of artificial intelligence (AI) and digital twin technologies is transforming the operational management of smart microgrids. AI-based models have shown high accuracy in forecasting renewable generation and load demand, optimizing energy dispatch strategies, and facilitating predictive maintenance [6,7]. Additionally, digital twins enable real-time simulation and validation of control strategies, allowing informed decisions under dynamic operating conditions [8]. However, unresolved challenges remain concerning the standardization of communication protocols, the development of interoperable control architectures, and the implementation of effective cybersecurity strategies. As microgrids increasingly depend on cloud-based management platforms, ensuring data privacy and protection against cyber threats becomes a critical concern [9,10]. Addressing these aspects is essential to encourage investment and support the global adoption of resilient, efficient, and economically sustainable microgrid solutions. To systematically analyze these technological trends and research gaps, this study applies the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology, ensuring a transparent and rigorous review process [11]. The aim is to consolidate the latest developments in smart microgrid management, focusing on energy storage technologies, AI-driven control strategies, and secure communication frameworks. This review offers actionable insights for researchers, industry professionals, and policymakers committed to advancing sustainable and intelligent energy systems.

The energy transition has driven a substantial rethinking of power system design and management, motivated by the need to reduce greenhouse gas emissions and improve energy security. Early studies primarily addressed the operational challenges associated with large-scale RES integration into conventional centralized grids, highlighting the volatility introduced by variable generation profiles and its implications for grid stability and reliability [1,12]. These investigations emphasized the importance of flexibility mechanisms and advanced forecasting techniques to maintain operational balance amid growing uncertainty. Subsequent research turned toward DERs and microgrid architectures as scalable approaches to mitigate these challenges [2,13]. In particular, the adoption of intelligent energy management systems (EMSs) enabled more dynamic and autonomous operation of decentralized grids. The integration of real-time monitoring platforms, including Supervisory Control and Data Acquisition (SCADA) and Advanced Metering Infrastructure (AMI), provided improved visibility and control over distributed assets, supporting the deployment of advanced demand response and optimization strategies [3,14,15]. Studies also recognized the relevance of hybrid AC/DC microgrid configurations, which facilitate the integration of diverse energy sources and storage technologies, offering greater operational flexibility and improved fault tolerance [4,16,17]. Previous reviews, such as Kamankesh et al., have addressed smart microgrid configurations with storage and renewable generation, providing broad insights into system architectures. In contrast, this study adopts a structured methodology to synthesize the latest advances in intelligent control, predictive analytics, and optimization technologies [18].

Progress in energy storage technologies has further strengthened microgrid resilience and performance. While BESSs remain widely adopted due to their fast response times and modular design, their economic competitiveness is constrained by high initial costs and limited service life [5,12,14]. To address long-duration storage requirements and seasonal variability, hydrogen-based storage solutions and pumped hydro storage (PHS) have been explored, each with distinct advantages and limitations [6,19,20,21]. Hybrid storage systems that combine BESS with hydrogen and PHS technologies have gained attention for their capacity to handle both short-term fluctuations and long-term balancing needs [7,22,23].

With the rapid evolution of digital technologies, AI, machine learning (ML), and digital twins have become central to microgrid control and optimization. AI-based forecasting methods, including deep learning and reinforcement learning algorithms, have demonstrated superior predictive accuracy for renewable generation and demand, enabling more precise energy dispatch [8,24,25,26]. Digital twins offer real-time virtual representations of physical systems, facilitating advanced scenario analysis, predictive maintenance, and resilience evaluations under diverse operational conditions [9,27,28,29]. The growing reliance on cloud-based platforms for microgrid management has also brought cybersecurity and data privacy to the forefront of current research. Several studies have raised concerns regarding system vulnerabilities to cyberattacks, underscoring the need for multi-layered security frameworks and standardized communication protocols [10,30,31]. Blockchain technologies are being examined as decentralized solutions for secure energy transactions and peer-to-peer trading platforms, providing integrity and transparency without centralized intermediaries [11,23,32]. The integration of AI-driven threat detection systems alongside blockchain platforms is viewed as a key approach to enhancing the security and resilience of future microgrids [33,34].

Alongside these technical developments, recent studies have explored the economic and financial aspects of microgrid deployment. The high capital requirements associated with advanced control platforms, energy storage, and renewable integration have led to the examination of alternative business models and financing schemes, including Energy-as-a-Service (EaaS), green bonds, and public–private partnerships [35,36,37]. Proposals to incorporate resilience metrics into financial assessments aim to quantify the long-term benefits of microgrid investments, particularly in the context of growing climate and operational risks [38,39]. These strategies are essential to support the scalability and long-term viability of microgrid projects, particularly in remote and underserved areas.

Despite these advancements, important research gaps remain, particularly regarding interoperability frameworks and the development of unified control architectures capable of integrating diverse technologies from multiple vendors [22,40]. Although explainable AI (XAI) and federated learning offer potential for enhancing transparency and privacy in AI-driven control systems, their practical implementation in microgrid environments is still limited and requires further investigation [41,42]. In response to these challenges, this study applies the PRISMA methodology [43], ensuring a transparent and rigorous process for identifying, selecting, and analyzing relevant scientific contributions. This structured approach consolidates recent advancements, identifies remaining research gaps, and proposes directions to promote the development of intelligent, resilient, and economically sustainable smart microgrids.

The analysis of the selected studies reveals that research is primarily concentrated on six interconnected thematic areas. These include the development of advanced EMSs and control architectures, highlighting the transition from centralized to distributed schemes and the growing focus on interoperability across heterogeneous systems. Another area is the integration of RES and hybrid microgrid configurations, which address technical and operational challenges associated with combining multiple generation technologies in complex hybrid AC/DC architectures. Significant efforts also focus on the deployment and optimization of advanced energy storage systems, examining both the technical performance and economic implications of hybrid storage solutions under varying demand scenarios. It is also important to recognize that the development and deployment of advanced energy storage systems and intelligent control strategies are strongly influenced by geographic and climatic conditions. Countries with high solar irradiance such as Australia, Spain, and parts of Latin America prioritize photovoltaic systems coupled with battery storage, while northern countries like Norway, Sweden, and Canada have focused on hybrid systems including hydro-based storage and seasonal strategies. Leading economies such as Germany, China, and the United States have developed robust energy storage and control infrastructures, often supported by progressive policies and long-term investments. These regional differences shape the technological priorities, research focus, and system-level integration strategies adopted across the globe. Therefore, understanding these contextual factors is crucial to interpreting the generalizability and applicability of the reviewed studies [44].

To meet these objectives, this paper is structured to provide a detailed analysis of current advancements and remaining challenges in smart microgrid management. This review provides a structured and thematic synthesis of recent advancements in smart microgrid management, focusing specifically on the integration of advanced energy storage systems (ESSs), intelligent control strategies, and optimization techniques. Unlike previous reviews, this study applies the PRISMA methodology to ensure transparency and replicability, and it incorporates a rigorous screening process based on five eligibility criteria. Moreover, it includes a keyword co-occurrence analysis to identify thematic clusters and current research directions. The reviewed studies, published between 2014 and 2025, reflect the most recent developments in the field. In contrast to earlier reviews that often address microgrids from a general or hardware-centric perspective, this paper emphasizes the convergence of control, prediction, and digitalization as pillars of next-generation microgrid systems. This review aims to provide a structured synthesis of recent advancements in the management and optimization of smart microgrids, with a particular focus on energy storage integration, intelligent control strategies, and predictive optimization techniques. To guide the analysis, the following research questions (RQs) were formulated: (1) What advanced energy storage configurations have been deployed in recent smart microgrid studies? (2) What types of control strategies—centralized, distributed, or adaptive—are most commonly implemented? (3) How are artificial intelligence, predictive analytics, and digital twins being used to support microgrid stability and performance? (4) What challenges and research gaps persist in achieving resilient, scalable, and secure smart microgrid operation? These questions shape the thematic structure and synthesis presented in the subsequent sections.

Following the introduction, Section 2 presents the applied methodology based on the PRISMA 2020 guidelines. Section 3 offers a comprehensive analysis of the findings, organized into six thematic areas: advanced EMSs and control architectures, integration of RES and hybrid microgrids, deployment and optimization of advanced storage systems, application of AI, predictive analytics, and digital twins, cybersecurity and privacy in energy management systems, and economic and resilience analyses. Finally, the paper concludes by synthesizing key findings and proposing future research directions to advance the development of intelligent, resilient, and economically sustainable smart microgrids.

2. Methodology for Systematic Literature Review

2.1. Introduction to PRISMA Methodology

This systematic review follows the PRISMA 2020 guidelines [45], which outline a structured approach for conducting literature reviews in technical areas such as renewable energy integration and smart microgrid management. PRISMA emphasizes transparent reporting and methodological consistency to support objective study selection and reduce bias.

While methodologies such as the Cochrane Handbook and MOOSE guidelines are widely used in medical fields, PRISMA provides flexible criteria appropriate for interdisciplinary domains combining engineering, technology, and socio-economic aspects. The process consists of four phases: identification, screening, eligibility and inclusion, and synthesis.

This review applies each phase with particular attention to reproducibility, given the diversity of study designs and topics in the smart microgrid domain. During the identification phase, relevant studies are retrieved using well-defined search strategies applied to selected academic databases. The screening phase involves reviewing abstracts to assess whether the studies meet predefined inclusion and exclusion criteria. In the eligibility and inclusion phase, full-text documents are evaluated to ensure that only studies aligned with the review objectives and methodological rigor are retained. Examples of this decision-making process, including reasons for exclusion based on relevance, clarity, or insufficient methodological development, are included in a supplementary annex. Finally, the synthesis phase integrates the selected studies to support this review’s conclusions.

Figure 1 illustrates the structured workflow applied throughout the selection process, from initial identification to final inclusion. Descriptions of each phase are provided in Section 2.2, Section 2.3, Section 2.4 and Section 2.5.

2.2. Identification Phase

The literature search was conducted in Scopus and Web of Science (WoS), selected for their academic rigor and coverage of peer-reviewed research in engineering, energy systems, and technology. The search was limited to English-language journal articles and conference papers published between 2014 and 2025.

Database-specific queries were formulated to ensure consistency and relevance. The exact expressions used are summarized in

Table 1. Search queries used for Scopus and Web of Science (WoS).

Scopus Query	WoS Query
TITLE-ABS-KEY ( “smart” AND “microgrid” AND “management” AND “modeling” ) AND PUBYEAR > 2013 AND PUBYEAR < 2026 AND ( LIMIT-TO ( LANGUAGE , “English” ) ) AND ( LIMIT-TO ( SRCTYPE , “j” ) OR LIMIT-TO ( SRCTYPE , “p” ) )	(ALL = (“smart”)) AND (ALL = (“microgrid”)) AND (ALL = (“management”)) AND (ALL = (“modeling”)) Refined by: Years 2014–2025, Language: English, Document Types: Article or Proceeding Paper.

. To minimize semantic ambiguity and ensure thematic precision, the search strings were deliberately limited to explicit and compound terms such as “smart”, “microgrid”, “management”, and “modeling”, thereby excluding abbreviations or loosely defined synonyms. This decision follows established PRISMA 2020 guidelines, which prioritize clarity, replicability, and alignment with clearly defined research questions.

The combined search yielded 597 records: 386 from Scopus and 211 from WoS. After removing 118 duplicates using bibliographic tools, 479 unique studies were retained for screening.

2.3. Screening Phase

The 479 documents retained after deduplication were screened to assess alignment with the scope of this review. Titles, abstracts, and metadata were evaluated against predefined criteria: publication period (2014–2025), English language, original research format (articles or conference papers), full-text availability, and thematic relevance to smart microgrid management, energy storage systems (e.g., BESS, hydrogen, pumped hydro), intelligent control, and optimization techniques.

Documents such as reviews, editorials, book chapters, and theses were excluded to focus the review strictly on primary research with validated methods and results.

This phase resulted in the selection of 346 studies—271 from Scopus and 75 from WoS—confirming broader coverage in the former. Figure 2 shows the distribution by year, highlighting increased research activity from 2016 onward, with peaks in 2022 and 2024. The upward trend continues in 2025 with 23 studies identified by midyear.

2.4. Eligibility and Inclusion Phase

Full-text evaluation was conducted for the 346 studies selected in the screening phase. To ensure methodological transparency and consistency, each study was scored using a structured matrix based on five predefined eligibility criteria: alignment with research objectives, methodological rigor, originality, data quality, and scientific influence.

To handle methodological heterogeneity across studies, the evaluation emphasized content relevance and internal consistency rather than imposing rigid formats. This allowed the inclusion of diverse modeling approaches, simulation frameworks, and control strategies, as long as they met the review’s thematic and analytical expectations.

Bias was mitigated through cross-review validation among authors. Disagreements were discussed until consensus was reached. A minimum threshold of 12 out of 15 points was required for inclusion. Additionally, a supplementary annex has been included to illustrate representative cases of inclusion and exclusion, based on the established eligibility criteria. This material supports transparency without overloading the main text.

Each criterion was rated on a three-level scale (1–3), as detailed below:

• Alignment with Research Objectives: Relevance to microgrid management, energy storage (e.g., BESS, hydrogen, PHS), AI-based control, and system optimization (1: Peripheral, 2: Related, 3: Highly Relevant).
• Methodological Rigor: Clarity in research design, model formulation, simulation, or validation (1: Needs Improvement, 2: Acceptable, 3: Strong).
• Originality and Technical Contribution: Introduction of new methods or configurations in smart microgrid operation (1: Minor, 2: Moderate, 3: Major).
• Data Quality and Analysis: Transparency, depth of results, and reproducibility (1: Satisfactory, 2: Good, 3: Excellent).
• Scientific Influence: Citation visibility and relevance in the field (1: Low, 2: Moderate, 3: High).

Figure 3 summarizes the scoring results. Studies falling below the minimum threshold were excluded not due to lack of merit, but because they did not meet the core analytical and thematic standards.

To enhance transparency, a supplementary annex (Appendix A.1) has been added with representative examples of included and excluded studies, including justifications based on the evaluation matrix.

A total of 66 studies, representing 19.1% of the evaluated works, met all criteria and were retained for synthesis.

2.5. Synthesis Phase

The synthesis phase included 66 studies that met the eligibility threshold, offering focused insights on energy storage integration, control strategies, and optimization in smart microgrid environments. Figure 4 summarizes their distribution by source and year. Most works (53) were published in peer-reviewed journals, led by IEEE Transactions on Smart Grid (10) and Applied Energy (8), followed by journals like Energies, Sustainable Energy Technologies and Assessments, and others with two or three entries each. The remaining 13 studies were published in conference proceedings, including ICEMS 2022, ECCE 2024, ISGT-Europe 2019, and AUPEC 2021. Chronologically, the number of relevant publications increased after 2016, peaking in 2018 and remaining consistently high through 2024. The year 2025, though ongoing, already contributes three studies. This set of works represents a balance between consolidated academic contributions and emerging lines of inquiry.

In addition to source distribution, Figure 4 includes a word cloud derived from the keywords of the selected studies, showing the dominant concepts across the reviewed literature. Terms such as energy management, microgrid control, renewable integration, energy storage, and optimization appear most frequently, reflecting the main technical concerns in this field. Based on this distribution, six thematic areas were defined to cluster the selected studies, enabling a structured synthesis and facilitating the discussion of research findings. Each thematic cluster contributes to addressing one or more of the research questions introduced in the study:

• Advanced Energy Management Systems and Control Architectures: This cluster directly addresses RQ2 by analyzing centralized and distributed control approaches, hierarchical frameworks, and the use of real-time management platforms such as SCADA and AMI.
• Integration of Renewable Energy Sources and Hybrid Microgrid Configurations: Focused on the technical complexity of integrating diverse generation sources, this area supports RQ4 by discussing architectural and operational strategies for managing system uncertainty and improving resilience.
• Deployment and Optimization of Advanced Energy Storage Systems: This cluster responds to RQ1 by reviewing various energy storage technologies and optimization strategies for improving energy balancing and operational flexibility in microgrids.
• Application of Artificial Intelligence, Predictive Analytics, and Digital Twins: This section contributes to RQ3, examining how predictive tools and AI-based models are used for forecasting, control optimization, and real-time simulation through digital twins.
• Cybersecurity and Privacy in Energy Management Systems: Aligned with RQ4, this theme explores methods to ensure secure communication, protect user data, and embed cybersecurity within the architecture of smart microgrids.
• Economic and Resilience Analysis of Microgrid Systems: This area also contributes to RQ4 by addressing trade-offs between cost and reliability, the design of resilient infrastructure, and investment strategies under operational uncertainty.

These clusters guide the discussion in the next sections, helping to organize the literature thematically and to illustrate how the selected studies contribute to answering the proposed research questions while identifying remaining gaps for future work.

3. Results and Discussion

While the reviewed studies collectively address major themes in smart microgrid development, notable differences emerge in methodological emphasis, control strategies, and contextual assumptions. For example, while centralized model predictive control (MPC) is praised for its precision in energy dispatch [13], several studies favor distributed or adaptive approaches due to their scalability and fault tolerance under uncertain conditions [19,46]. Similarly, although hybrid energy storage systems combining batteries and supercapacitors offer performance advantages [15], they also present higher integration and cost complexity [47,48]. There is also divergence in the application of artificial intelligence: some studies report performance improvements using reinforcement learning [49], while others highlight overfitting risks or interpretability challenges [42]. In addition, although many papers advocate for interoperability and standardization, few provide actionable frameworks or validated protocols [20]. These inconsistencies reveal critical knowledge gaps, particularly in the practical implementation and scalability of intelligent microgrid solutions across varied geographical and regulatory environments. This synthesis aims not only to categorize trends but also to highlight such contrasts as a foundation for future targeted research.

3.1. Advanced Energy Management Systems and Control Architectures

Energy management in smart microgrids has gained importance due to the growing complexity of electrical networks and the integration of DERs and RES. This section reviews current developments in control architectures, real-time monitoring platforms, demand response strategies, and the shift toward decentralized and automated management systems.

3.1.1. Architectures for Energy Management: Centralized and Distributed Approaches

Initial research concentrated on centralized control architectures [12,13], where decision-making processes were conducted from a central control center. However, the proliferation of DERs and the increasing demand for resilience have driven a shift toward distributed control models [14,15], which support local, real-time decision-making and improve the system’s responsiveness to disturbances.

Several studies have proposed hierarchical control schemes to efficiently manage the different operational layers of microgrids [24,26]. These architectures combine centralized supervisory control with local and field-level subsystems, optimizing network stability and economic performance [25,50]. The literature also highlights the importance of effective control mechanisms to mitigate the uncertainty caused by the variability of RES [41,51,52].

Despite these advancements, interoperability challenges among devices from different manufacturers and the absence of standardized communication protocols remain unresolved barriers [19,53]. Several studies emphasize the need to develop common frameworks that facilitate the integration of heterogeneous technologies [20,54,55]. In this context, the implementation of intelligent agents and multi-agent systems has been identified as an effective solution for distributed management, enhancing system adaptability to unforeseen events and operational contingencies [56,57].

3.1.2. Real-Time Monitoring and Control Platforms (SCADA and AMI)

Real-time monitoring in microgrids has advanced through the deployment of SCADA and AMI platforms [16,17], enabling continuous observation of key system variables. These platforms offer real-time visualization of operational conditions and support decision-making based on updated system data [47,58,59]. The integration of AI has further expanded their capabilities, allowing these systems to anticipate possible failures rather than simply record historical information [22,60].

The use of digital twins has also become an important tool for creating virtual models of physical systems, enabling the simulation of complex scenarios and analysis of contingency plans without affecting actual operations [27,29,31]. The implementation of these technologies requires secure and stable communication infrastructures to preserve data integrity and system availability, minimizing exposure to cybersecurity risks [28,42,61]. Recent studies have explored the application of embedded systems in distributed environments to enhance local control and enable remote monitoring functions [32,62].

3.1.3. Demand Response and Predictive Control Strategies

Active demand management has become a central focus in microgrid operation, promoting user participation in optimizing energy consumption [39,43,63]. demand response (DR) programs enable flexible adjustment of consumption patterns, contributing to the balance of supply and demand in systems with high RES penetration [64,65,66]. The application of model predictive control (MPC) is frequently identified as an effective approach for implementing predictive control strategies [67,68]. These models forecast system behavior and optimize energy dispatch decisions while considering technical constraints, economic factors, and user comfort requirements [69,70,71].

However, the lack of high-quality datasets and concerns over user data privacy present significant barriers to the effective deployment of these strategies [72,73,74]. The use of deep learning and neural network models is gaining attention for enhancing prediction accuracy, although these methods require substantial data resources and computational capacity [75,76,77].

3.1.4. Trends Toward Decentralized Intelligence and Interoperability Standards

Recent advancements in microgrid management emphasize the adoption of decentralized intelligence architectures, where energy resources operate autonomously and coordinate through local optimization algorithms [49,78]. Reinforcement learning and adaptive control strategies support continuous performance improvements in microgrid operations [12,13]. In parallel, growing interest exists in the development of open platforms and interoperability standards to enable secure and efficient communication among devices from different manufacturers [14,79]. The establishment of such standards is critical to facilitate microgrid participation in local energy markets and peer-to-peer energy trading schemes [15,80,81].

Future research should prioritize the implementation of explainable AI (XAI) models to improve transparency in AI-based control systems, and the exploration of blockchain technologies to support secure and decentralized energy transaction management [30,58,67]. Furthermore, the creation of open data repositories is essential to enhance the training of advanced models while preserving user privacy and ensuring data security [22,55,61].

3.2. Integration of Renewable Energy Sources and Hybrid Microgrid Configurations

The integration of RES into microgrids is a fundamental enabler for developing sustainable and low-emission energy systems. However, the variability and unpredictability of renewable generation, combined with the increasing complexity of hybrid AC/DC microgrid architectures, introduce considerable technical and operational challenges. This section reviews the latest research and technological developments aimed at addressing these issues.

3.2.1. Challenges of Renewable Integration and Variability Mitigation

The integration of RES such as solar photovoltaic and wind power into microgrids introduces considerable variability, posing challenges to system stability and reliability [12,15,79]. These issues become more pronounced in configurations with high renewable penetration, where generation fluctuations must be rapidly compensated to maintain power quality [80,81]. Several studies emphasize the importance of advanced forecasting models and energy management systems to mitigate the effects of intermittency [24,26]. Moreover, the absence of standardized methodologies for modeling and predicting the dynamic behavior of RES within microgrids complicates the design of effective control strategies [25,51,82].

Hybrid solutions that combine RES with advanced energy storage systems and DR programs have been proposed to smooth out generation fluctuations and ensure continuous energy supply [41,52,53]. However, the economic viability of these solutions remains a significant barrier, particularly in remote or islanded microgrid environments where implementation costs are higher and financial resources are limited [21,54,57].

3.2.2. Hybrid AC/DC Microgrid Configurations

Hybrid AC/DC microgrids have emerged as a viable alternative to overcome the limitations of conventional AC-only architectures, offering enhanced flexibility for integrating diverse energy sources and storage technologies [16,17,83]. These configurations support efficient power conversion and distribution, reducing energy losses and improving overall system resilience [36,58].

Studies highlight that adopting hybrid configurations facilitates the integration of DERs, including photovoltaics, wind turbines, and battery storage systems [22,47,60,84]. Additionally, hybrid microgrids offer operational advantages in load balancing, fault tolerance, and the simultaneous support of AC and DC loads with improved efficiency [23,61]. Despite these benefits, challenges remain regarding control coordination and the development of robust protection schemes, which require further investigation and standardization efforts [34,48,62].

3.2.3. Control and Coordination of Hybrid Microgrids

Reliable control and coordination strategies are essential for the stable operation of hybrid AC/DC microgrids under varying generation and load conditions [39,43,66]. Both centralized and decentralized control approaches have been explored, with decentralized architectures providing greater scalability and resilience, particularly in systems with high renewable penetration [70,72]. Advanced control strategies based on hierarchical frameworks integrate multiple control layers, ranging from primary voltage and frequency regulation to higher-level optimization of energy dispatch and economic performance [75,76]. The implementation of predictive control models and AI-based algorithms is also gaining prominence, enabling real-time optimization and adaptive control in complex hybrid systems [49,77]. Ensuring the safe and stable operation of smart microgrids is a critical concern, particularly as system complexity increases with the integration of distributed energy resources, intelligent controllers, and communication networks. Operational safety involves coordinated protection schemes, fault detection and isolation, and control redundancy to prevent cascading failures. Stability is also influenced by the dynamic behavior of power electronics and the responsiveness of control systems under disturbances. Moreover, with the increasing use of digital communication and AI-based control, cybersecurity has become an essential aspect of operational resilience. Recent studies have emphasized the need for fault-tolerant control architectures, secure communication protocols, and real-time monitoring systems to guarantee reliable operation under both normal and abnormal conditions [85].

Nonetheless, achieving seamless coordination among diverse energy sources, storage systems, and controllable loads remains a key challenge. The development of interoperable communication protocols and standardized control interfaces is critical to facilitate effective interaction between heterogeneous system components and to enhance the reliability and efficiency of hybrid microgrid operations [22,48,61].

3.2.4. Future Role of Sector Coupling and Energy Market Participation

The future integration of RES in microgrids is expected to increasingly involve sector coupling, connecting the electricity sector with heating, cooling, and transportation systems to improve energy efficiency and support decarbonization efforts [36,58,83]. Sector coupling strategies enable flexible utilization of surplus renewable generation, such as directing excess electricity to electric vehicle charging or thermal storage systems, thereby enhancing overall system efficiency [47,60].

Additionally, microgrid participation in local and regional energy markets represents a promising opportunity for future development. Several studies emphasize the importance of establishing regulatory frameworks that enable microgrids to provide ancillary services and participate in peer-to-peer energy trading schemes [22,23,48]. Blockchain technology and decentralized transaction platforms are also being explored as mechanisms to facilitate transparent and secure energy trading within microgrid communities [34,62].

Future research should focus on developing business models that support sector coupling and market participation while ensuring the financial sustainability and operational resilience of microgrids [39,43,66]. These efforts should be complemented by the creation of advanced optimization tools capable of assessing the economic and environmental impacts of sector coupling strategies under various operational scenarios [70,72,77].

3.3. Deployment and Optimization of Advanced Energy Storage Systems

The integration of advanced energy storage systems (ESSs) is fundamental to enhancing the flexibility, reliability, and efficiency of smart microgrids. With the increasing penetration of RES, storage technologies play a critical role in balancing supply and demand, mitigating intermittency, and maintaining grid stability. This section reviews recent developments in storage technologies, including comparative analyses, optimal sizing and placement strategies, lifecycle cost evaluations, and emerging trends toward hybrid storage solutions and circular economy practices. Adaptive control strategies are becoming increasingly relevant in microgrid applications, particularly as the integration of renewable energy sources introduces high levels of uncertainty and volatility. Unlike static or pre-programmed methods, adaptive approaches can adjust their control parameters in real time based on environmental inputs, system performance, or changing grid conditions. This capability enhances system resilience, improves energy dispatch decisions, and enables better response to faults or operational disturbances. Recent studies have explored adaptive fuzzy controllers, model predictive control with online tuning, and reinforcement learning-based controllers as effective solutions for dynamic and uncertain environments [46].

3.3.1. Comparative Analysis of Storage Technologies (BESS, Hydrogen, PHS)

Among the available storage solutions, BESSs have received significant attention due to their high energy density, rapid response capabilities, and modular deployment options [12,14,15]. Despite these technical advantages, challenges persist regarding high capital costs, limited service life, and environmental concerns associated with battery disposal [26,41,82].

Hydrogen-based storage systems are gaining interest as a long-term solution, particularly for large-scale and seasonal storage applications [19,21]. These systems offer substantial energy capacity and the potential for integration with power-to-gas technologies; however, efficiency losses and the requirement for advanced hydrogen infrastructure present notable challenges [20,55].

Pumped hydro storage (PHS) remains one of the most mature and cost-effective technologies for large-scale energy storage [16,17]. PHS provides excellent long-term storage capabilities and high efficiency but is geographically constrained, limiting its deployment in certain regions [36,83].

The comparative analysis of these technologies underscores the importance of selecting appropriate storage solutions based on the specific operational requirements and geographic conditions of each microgrid.

3.3.2. Optimal Sizing and Placement of Energy Storage

Optimizing the sizing and placement of ESSs is essential to maximize their economic and technical benefits. Various studies propose advanced optimization algorithms, including genetic algorithms and mixed-integer linear programming, to determine the ideal capacity and location of storage units within microgrids [22,47,86]. Properly sized storage systems enhance grid stability, reduce operational costs, and minimize energy curtailment [23,27]. Additionally, optimal placement supports efficient energy distribution, reduces transmission losses, and facilitates local voltage and frequency regulation [48,61].

Multi-objective optimization approaches have gained prominence, balancing technical objectives such as reliability and efficiency with economic considerations, including investment and operational costs [34,42]. However, the real-world application of these models faces challenges related to computational complexity and the availability of accurate demand and generation data [43,70].

3.3.3. Lifecycle Cost and Performance Evaluation

Evaluating the lifecycle costs and performance of storage technologies is essential for guiding investment decisions and planning long-term system operation. Studies emphasize the need to account for capital expenditures, operation and maintenance costs, performance degradation over time, and end-of-life disposal requirements [77,78]. Analyses indicate that although BESSs currently involve higher initial costs, technological developments are gradually improving their economic competitiveness [12,14].

Hydrogen-based storage systems, despite requiring considerable initial investments, are being examined for their potential in long-duration storage applications and integration with sector coupling strategies [19,20]. PHS remains a cost-competitive option for large-scale energy storage over extended periods [16,36]. However, incorporating lifecycle performance indicators into investment evaluations continues to be a complex task, highlighting the need for consistent and standardized assessment methodologies [22,34].

3.3.4. Research on Hybrid Storage Systems and Circular Economy Approaches

Recent research underscores the growing interest in hybrid energy storage systems, which combine multiple technologies to capitalize on their complementary characteristics [22,47]. For instance, coupling fast-response BESSs with high-capacity hydrogen or PHS systems enables microgrids to manage both short-term fluctuations and long-term energy balancing requirements effectively [23,27].

In parallel, circular economy strategies are being actively explored to minimize the environmental impact of storage technologies. These include battery recycling programs, second-life applications for retired batteries, and the design of modular, easily recyclable storage units [48,61].

Future research should focus on developing integrated frameworks that combine hybrid storage optimization with circular economy principles, ensuring both economic and environmental sustainability [34,42]. Establishing policy incentives and regulatory frameworks to support the deployment of hybrid systems and promote sustainable materials management will be critical for scaling advanced energy storage solutions in microgrid environments [43,70].

3.4. Application of Artificial Intelligence, Predictive Analytics, and Digital Twins

The increasing complexity of smart microgrid operations and the integration of high shares of RES have accelerated advancements in the application of AI, predictive analytics, and digital twins. These technologies support improved decision-making, enhance system reliability, and enable predictive maintenance and operational optimization.

3.4.1. AI-Based Forecasting and Dispatch Optimization

Accurate forecasting of renewable generation and load demand is critical for optimizing microgrid operations. AI-based models, including machine learning (ML) and deep learning techniques, have demonstrated superior performance in handling complex forecasting tasks under uncertain and dynamic conditions [12,13,79]. Studies report successful applications of neural networks, support vector machines, and reinforcement learning algorithms for short-term load and renewable generation forecasting, achieving higher accuracy compared to traditional statistical models [14,80,81].

AI-driven dispatch optimization strategies contribute to more efficient energy management by minimizing operational costs and maximizing the utilization of RES [24,25,26]. Furthermore, hybrid models that combine AI methods with metaheuristic algorithms, such as genetic algorithms and particle swarm optimization, have been developed to improve dispatch efficiency in multi-objective optimization scenarios [50,51,82]. Despite these advances, challenges related to model interpretability and limited data availability persist [41,52,53].

3.4.2. Anomaly Detection and Fault Diagnosis Using Machine Learning

ML techniques have proven effective in anomaly detection and fault diagnosis within microgrid environments, supporting preventive maintenance and reducing operational costs [19,20,21]. Supervised learning models are frequently applied for fault classification, while unsupervised learning approaches help detect unknown anomalies based on operational patterns [54,55,56].

Recent studies highlight the use of convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to analyze time-series data and identify critical failures in real time [16,17,57]. Ensemble learning techniques, including Random Forest and Gradient Boosting, further enhance the accuracy and robustness of fault diagnosis frameworks [36,83]. However, the effectiveness of these models depends heavily on access to large, representative datasets, which remains a significant limitation [58,59,60]. Addressing this challenge requires the development of data-sharing frameworks and federated learning models that enable distributed data utilization while preserving privacy [22,31,87]. Recent studies highlight the use of convolutional neural networks (CNNs) and deep reinforcement learning models to identify optimal energy management strategies in smart microgrids operating under uncertainty [88].

3.4.3. Development and Implementation of Digital Twins

Digital twins are increasingly employed to model and simulate the dynamic behavior of microgrid systems, providing virtual environments for testing control strategies and evaluating system performance under diverse scenarios [23,27,28]. These virtual replicas facilitate real-time monitoring and predictive maintenance, enhancing operational efficiency and reducing downtime [42,48,61].

Research demonstrates that integrating AI-driven analytics with digital twins improves their predictive capabilities, enabling more accurate failure forecasting and optimized maintenance scheduling [30,34,62]. Digital twins also play a critical role in supporting scenario-based planning for system expansion and resilience assessments [32,39].

Despite these benefits, challenges remain in developing standardized frameworks that ensure interoperability and data consistency across different platforms and microgrid components [43,63,64]. Future research should prioritize the creation of open-source digital twin platforms and explore their integration with decentralized control architectures [65,66,67].

3.4.4. Explainable AI and Data Sharing Frameworks for Collaborative Learning

As AI models become more advanced and integrated into decision-making processes, the need to improve their transparency and interpretability has become increasingly relevant [69,70,73]. Explainable AI (XAI) techniques address this by clarifying how AI-based decisions are made, particularly in optimization and control tasks within microgrids [74,75,76].

Collaborative learning approaches, such as federated learning, offer alternatives for model development while protecting sensitive data [49,77,78]. These methods allow multiple microgrids or energy management systems to train models collectively without exchanging raw data.

Future efforts should focus on combining XAI techniques with collaborative learning frameworks to achieve transparent, secure, and efficient AI deployment in microgrids. Establishing common standards for data sharing and model evaluation will support cooperation between research institutions and industry organizations [34,42,48].

3.5. Cybersecurity and Privacy in Energy Management Systems

The digitalization of smart microgrids has improved operational flexibility but has also introduced new cybersecurity and data privacy risks. As microgrids become more dependent on interconnected platforms, real-time data exchange, and distributed control systems, protecting these infrastructures from cyber threats is now a research priority.

3.5.1. Emerging Cyber Threats and Vulnerabilities in Microgrids

The increasing complexity and connectivity of modern microgrids have expanded the range of security vulnerabilities that can be exploited through various attack methods [15,31]. These include unauthorized access, data breaches, denial-of-service attacks, and malware targeting energy management platforms, all of which disrupt operations and compromise information integrity.

The use of third-party platforms and cloud services for energy management increases exposure to data leaks and operational failures [23,32]. These risks affect both the protection of information and the stable operation of the energy supply to end users.

3.5.2. Secure Communication Protocols and Privacy Techniques

To reduce cybersecurity risks, several communication protocols and privacy protection methods have been proposed. Encryption protocols, including Transport Layer Security (TLS) and Internet Protocol Security (IPsec), help secure data exchanges between DERs and control centers [15,23].

Additionally, privacy protection strategies such as differential privacy and homomorphic encryption allow the analysis of aggregated data without exposing individual consumption patterns or sensitive operational information [31,32].

3.5.3. Multi-Layered Security Frameworks and Standards

An effective cybersecurity strategy for microgrids requires the deployment of multi-layered security frameworks that address vulnerabilities across all network levels, from physical infrastructure to application-level controls [15,31]. These frameworks typically include device authentication, network segmentation, intrusion detection systems, and encryption protocols.

Standardization organizations have published cybersecurity guidelines specifically designed for energy infrastructures, including the IEC 62443 series and the NIST Cybersecurity Framework [23,32]. However, the implementation of these standards varies across regions and system designs, highlighting the need for greater regulatory consistency and enforcement.

3.5.4. AI-Based Threat Prediction and Blockchain for Secure Transactions

The application of AI for threat prediction focuses on improving the resilience of energy management systems. AI models analyze network traffic, identify anomalies, and predict cyberattacks before they occur, enabling preventive defense strategies [15,31].

At the same time, blockchain technology is being explored as a decentralized method to secure energy transactions in microgrids [23,32]. By recording transactions in an immutable and transparent manner, blockchain facilitates secure peer-to-peer energy trading without reliance on central intermediaries.

Future research should examine how AI-based threat detection can be integrated with blockchain platforms to develop automated cybersecurity solutions for microgrids. Defining standardized protocols for secure data exchange and transaction verification will be essential to support reliable and scalable deployment of these technologies.

3.6. Integration of Smart Microgrids with Virtual Power Plants (VPPs)

Virtual Power Plants (VPPs) are emerging as a complementary and scalable paradigm for managing distributed energy resources. Although VPPs differ from microgrids in structure and scope, they share several operational functions, such as distributed energy aggregation, peak shaving, demand response, and ancillary services. Recent studies suggest that microgrids can act as functional subsystems within VPP frameworks, enabling coordinated scheduling and market participation. In this context, microgrids provide local autonomy and resilience, while VPPs offer centralized control and market interfacing. This convergence is particularly relevant for the deployment of advanced control strategies and intelligent energy management systems, enabling dynamic interaction with grid operators and virtual aggregators. The incorporation of VPP-related functionality into microgrid design further reinforces their strategic role in the future of decentralized and transactive energy systems [89,90].

3.7. Economic and Resilience Analysis of Microgrid Systems

Economic viability and resilience are essential considerations for the successful deployment and long-term operation of microgrid systems. While microgrids improve energy reliability and support the integration of RES, their development depends on achieving financial sustainability and design strategies capable of managing operational risks.

3.7.1. Investment Analysis and Financial Viability

Investment planning for microgrid systems requires detailed evaluation of both initial capital costs and long-term operational benefits [13,24]. Although advanced technologies such as ESSs and intelligent control platforms require considerable upfront investment, these costs can be balanced by lower operating expenses and improved energy management over time [19,55].

Economic analyses increasingly apply life cycle cost assessments to compare technology options and their expected financial outcomes over extended periods [17,37]. Policy incentives and government subsidies also influence the financial attractiveness of microgrid initiatives, particularly in remote areas and developing regions [30,36].

However, regulatory uncertainty and fluctuating energy market prices introduce financial risks that complicate investment planning [32,39]. These challenges highlight the importance of developing economic models that incorporate sensitivity analyses and risk evaluations to support informed decision-making.

3.7.2. Operational Cost Reduction and Efficiency Improvement

Lowering operational costs and improving the efficiency of microgrid systems are essential to support their long-term viability [68,70]. Recent studies emphasize the importance of advanced energy management platforms in reducing energy losses, optimizing generation dispatch, and decreasing dependence on fossil fuel-based backup systems [71,75].

The adoption of predictive maintenance strategies and real-time monitoring platforms also helps prevent equipment failures and extend the service life of critical infrastructure [13,24]. Demand response programs contribute by reducing peak load demands, which lowers operational costs related to capacity constraints and external grid support [19,55]. Advanced optimization algorithms, including multi-objective approaches, are applied to balance economic performance and technical reliability, ensuring that microgrids operate efficiently under varying load and generation conditions [17,37].

3.7.3. Resilience Assessment and Design Strategies

Resilience has become an important performance indicator for microgrids, particularly in response to risks associated with natural disasters, cyber threats, and supply chain disruptions [30,36]. Resilient microgrid designs incorporate redundancy, adaptive control mechanisms, and decentralized energy resources to maintain continuous operation during adverse events [32,39].

Studies propose resilience metrics based on system recovery time, the ability to supply critical loads during disruptions, and the effectiveness of operational recovery strategies [68,70]. Simulation tools and scenario-based analyses are frequently used to assess resilience under various failure and recovery scenarios [71,75]. Design strategies centered on modularity and scalability further enhance resilience by allowing system reconfiguration and phased upgrades without major infrastructure changes [13,24].

3.7.4. Business Models for Resilient Microgrids and Financing Mechanisms

Developing sustainable business models is essential to support the deployment of resilient microgrids. Models such as Energy-as-a-Service (EaaS) and microgrid leasing options help reduce upfront financial barriers for end-users while providing professional management and maintenance services [19,55].

Financing mechanisms including green bonds, public–private partnerships, and performance-based contracts are increasingly used to support microgrid projects [17,37]. These models improve access to capital, reduce financial risks, and enable monetization of resilience and ancillary services provided by microgrids [30,36]. Blockchain platforms and decentralized marketplaces also introduce new possibilities for peer-to-peer energy trading and dynamic pricing models that encourage efficient energy use and investments in system resilience [32,39].

Future research should explore how these business models can be aligned with regulatory frameworks to support sustainable and scalable microgrid development [68,70].

Table 2. Consolidated research findings and future research directions in smart microgrids.

Topic	References	Research Challenges and Future Directions
Advanced Energy Management Systems and Control Architectures	[12,13,14,15,24,26]	S-037 explores hierarchical control for solar-storage systems. S-127 proposes edge-based architectures for real-time decision-making. S-169 compares rule-based with AI-enhanced control strategies. Challenges include establishing standardized communication protocols for interoperability; integrating blockchain to ensure secure decentralized coordination; adopting explainable AI models in multi-agent architectures; and enhancing adaptive distributed intelligence using MAS. [19,20,22,30,56,57,67]
Integration of Renewable Energy Sources and Hybrid Microgrid Configurations	[12,15,79,80]	S-037 highlights PV-battery integration with hybrid inverters. S-059 evaluates control in hybrid AC/DC topologies. S-299 discusses energy sector coupling. Challenges involve establishing protection and interoperability standards for AC/DC microgrids; expanding sector coupling with heating/cooling and EVs; and developing blockchain-enabled peer-to-peer markets. [34,48,62,83]
Deployment and Optimization of Advanced Energy Storage Systems	[12,14,15,19,20,21]	S-169 presents optimization of hybrid battery-supercapacitor ESS. S-232 reviews lifecycle models for storage. S-105 investigates predictive control for ESS dispatch. Challenges include improving cost/lifecycle of hybrid systems; adopting circular economy principles; and applying multi-objective optimization frameworks for storage sizing and location. [22,43,47,48,61,70]
Application of Artificial Intelligence, Predictive Analytics, and Digital Twins	[12,13,19,21,79]	S-124 implements federated learning in EMSs. S-136 integrates predictive maintenance with ML. S-192 explores digital twin platforms. Challenges include enhancing privacy through federated AI; standardizing open-access digital twin models; and improving anomaly detection with deep learning. [21,27,49]
Cybersecurity and Privacy in Energy Management Systems	[15,23,31,32]	S-232 outlines a multi-layer cybersecurity framework. WoS-086 applies blockchain for secure data exchange. WoS-066 discusses regulatory gaps in data privacy. Challenges include real-time threat detection using AI; harmonizing data standards and legal compliance; and integrating cybersecurity into all microgrid layers. [15,31,32]
Economic and Resilience Analysis of Microgrid Systems	[13,17,19,24,55]	S-209 models resilience in grid-tied microgrids. S-127 evaluates cost-benefit of microgrid expansion. S-270 introduces Energy-as-a-Service models. Challenges include robust financial models under uncertainty; innovative financing mechanisms like green bonds; and embedding resilience indicators in economic assessments. [30,37,39]

summarizes the main findings and research gaps identified in this review. For each technological area, the corresponding studies are cited, and current challenges and future research directions are outlined.

In addition to the consolidated findings presented in Table 2, a more detailed comparative analysis of selected representative studies is provided in

Table 3. Comparative analysis of key studies in smart microgrid research.

Reference	Methodology/Model	Main Contribution	Limitations/Gaps
[13]	MPC for energy dispatch	High dispatch precision in MGs	Low flexibility in centralized schemes
[19]	Adaptive distributed control	Scalable/fault-tolerant under uncertainty	Requires real-time communication
[15]	Hybrid ESS optimization	Operational metrics for hybrid BESS	Integration/cost complexity
[49]	Reinforcement learning (RL)	Improved fault detection via AI	Risk of overfitting, low explainability
[47]	Battery–SC hybrid storage	Efficiency improvements in BESS	Complex control and maintenance
[48]	Lifecycle-based storage planning	Multi-objective sizing/placement	Scalability in dynamic conditions
[20]	Standardization protocols	Highlighted comm. limitations in MGs	Lack of validated real deployments
[42]	Explainable AI (XAI)	Reliable and transparent diagnostics	High computational burden
[30]	Blockchain for trans. security	Secure decentralized energy exchange	Unclear regulatory standards
[34]	P2P energy trading platforms	Enabling distributed market models	Scalability/user engagement issues

. This table outlines the methodologies applied, the key contributions made by each study, and their reported limitations or knowledge gaps. This analytical matrix enhances the paper’s depth by highlighting not only what each study contributes but also where further research is needed.

In conclusion to Section 3, the findings discussed across the thematic areas directly address the four research questions presented in the introduction. Hybrid storage architectures (RQ1) have been widely implemented in recent studies to enhance flexibility and responsiveness. Control approaches (RQ2) range from centralized methods to adaptive multi-agent systems, as seen in multiple references. The integration of predictive tools and AI-based strategies (RQ3) has become increasingly prevalent, with applications such as reinforcement learning, digital twins, and deep learning models. Finally, persistent challenges (RQ4) include secure data exchange, lack of standardized AI deployment practices, and resilience coordination under uncertainty.

4. Conclusions

This review synthesized the findings of 66 selected studies focused on smart microgrid management, organizing the analysis around four guiding research questions (RQ1–RQ4). Each question helped structure the review and identify areas of convergence as well as limitations in the current body of knowledge.

For RQ1, which examines energy storage configurations, most studies explored hybrid approaches involving batteries, hydrogen storage, and pumped hydro systems. While some of these configurations show technical viability in simulation environments, their deployment remains constrained by cost–performance trade-offs, particularly in isolated or infrastructure-limited regions. Further work should examine how these configurations behave under variable load profiles and develop multi-criteria decision models that account for degradation, spatial constraints, and resource availability.

RQ2 concerns control strategies. The findings include both centralized techniques such as model predictive control, and distributed schemes including agent-based models. Some studies demonstrated potential in combining these approaches within layered control architectures. However, technical challenges persist regarding convergence speed, fault tolerance, and consistency in distributed decision-making. More experimentation is needed with real-time hardware emulation and uncertainty quantification in these control layers.

Regarding RQ3, which involves the use of artificial intelligence and predictive models, tools such as reinforcement learning, deep neural networks, and hybrid forecasting techniques were frequently employed. These models were applied to areas such as load prediction, energy dispatch, and anomaly detection. Nonetheless, limitations appear in model transparency, generalizability, and training data availability. Current efforts would benefit from applying explainable AI principles and evaluating algorithmic performance on standardized open datasets shared under privacy-preserving frameworks.

RQ4 addressed challenges that affect scalability, reliability, and data security in smart microgrid environments. Issues identified include fragmented cybersecurity practices, lack of interoperability standards, and incomplete frameworks for assessing operational resilience. Ongoing research would gain from defining minimum acceptable thresholds for secure communication, exploring recovery-oriented design metrics, and validating coordination protocols in mixed-vendor system architectures.

Beyond addressing the current gaps, the review identified a set of precise research questions and unresolved engineering problems that can guide further investigations:

• What storage combinations deliver the best trade-off between short-term response time and long-duration capacity under constrained capital budgets?
• How can distributed control agents resolve local conflicts without requiring central coordination or frequent communication in resource-limited networks?
• Which methods allow real-time AI models to provide interpretable feedback for operational decisions in dynamic grid conditions?
• What is the minimum communication infrastructure required to maintain secure system operation when integrating blockchain or AI-based monitoring?
• How can financial planning tools incorporate risk-adjusted cost models that respond to regulatory change and resilience expectations?

While the reviewed studies partially address the research questions, they also reveal several technical and methodological limitations. Closing these gaps will require targeted experimentation, model validation in operational settings, and interdisciplinary collaboration across control engineering, computer science, and regulatory planning.