A systematic literature review on the use of digital twin technology in electrical engineering research and applications

Introduction

Due to its ability to produce real-time simulation, enable predictive maintenance, and enhance system monitoring, digital twin (DT) technology has seen increased growth in engineering disciplines, and with it, a clear pathway toward advancing the design and operation of complex electrical systems (Ahmadi et al., 2021; Mansour et al., 2023). Although a rise in DT research in electrical engineering has been documented, it has been relatively slow compared to other disciplines, such as manufacturing, aerospace, and smart cities (Pan et al., 2020; Chen et al., 2022). Furthermore, effectively applying DTs in engineering curricula remains a challenge, due to technical complexity and resource requirements, which limit the widespread adoption of DT-enabled learning (Damaševičius and Zailskaite-Jakšte, 2024; Ezeoguine and Kasumu, 2024). Our literature review looks at trends and current knowledge in research on DT applications in electrical engineering. The results of this review provide a clear snapshot of the key areas in DT engineering, how these areas are connected, the major gaps in existing literature, and priorities for future research. These findings help build awareness and encourage the incorporation of DTs into engineering education by supporting more effective approaches to instruction, learning experiences, and system development using DT technologies.

The literature on DTs encompasses research pertaining to a broad range of applications, including manufacturing, healthcare, infrastructure, and energy systems. However, this review specifically targets electrical engineering applications. We restricted our systematic literature review to DT applications within electrical engineering in recognition of the growing number of DT implementations in these fields, particularly in power systems, embedded systems, electric machines, and IoT-enabled platforms (Mansour et al., 2023; Khalyasmaa et al., 2023). We excluded research related to non-engineering or non-digital twin-focused cyber-physical systems, as much of the recent literature has focused on broader conceptual discussions or applications outside the scope of electrical engineering. Research related to purely virtual simulations without bidirectional physical-digital coupling was also excluded, as such systems do not fully align with the DT paradigm of synchronized virtual and physical entities.

This study answers the research question: What is the current state of research on DT applications in electrical engineering? Our results are presented in two parts. First, we describe the selected articles using key characteristics, including publication year, publication venue and type, first author's country affiliation, engineering disciplines represented, theoretical or conceptual frameworks, research design, sampling approach and sample size, and the study context. Second, we synthesize the literature by identifying and organizing the major research themes that emerge across the studies. These findings help build awareness and encourage the incorporation of DTs into engineering education by supporting more effective approaches to instruction, learning experiences, and system development using DT technologies.

Methods

In this study, we followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework to structure and document the Systematic Literature Review (SLR) process. PRISMA outlines three core stages: (1) identification, where relevant studies are gathered from multiple databases using targeted search terms; (2) screening, where articles are reviewed at the abstract and full-text levels based on predefined exclusion criteria; and (3) inclusion, where the final set of eligible studies is examined in depth to answer the research questions. PRISMA was chosen because it offers a clear, transparent, and methodologically robust approach for conducting SLRs. The review process began by applying several keyword combinations across a range of academic databases. We used a total of three search terms/phrases in this study: “digital twin” AND “electrical engineering”, “digital twin” AND (“power system”OR “smart grid” OR “signal processing”), and “digital twin” AND (“embedded system^*” OR “VLSI” OR “chip design”). The databases used to find articles were Google Scholar, Web of Science, IEEE Xplore Digital Library, ERIC, Compendex (Engineering Village), ScienceDirect, and Wiley Online Library.

Data collection

In Figure 1, we present the complete article selection process. We excluded any articles that met one or more of the following seven exclusion criteria (EC), as they were not aligned with the scope and objectives of this study: EC1) published before 2015; EC2) not directly related to electrical engineering disciplines; EC3) not focused on DT technology or not encompassing the full scope of cyber-physical systems integration; EC4) work-in-progress abstracts or short papers EC5) published in a language other than English; EC6) not open-access or unavailable for full-text retrieval; or EC7) focused on engineering disciplines outside the strict scope of electrical (e.g., mechanical, biomedical, or civil engineering applications without a clear EE component).

As illustrated in Figure 1, a total of 540 articles were identified across six databases using the combination of three search strings. To complete the identification stage, the first author first removed duplicate entries and then conducted an initial review of abstracts, followed by full-text screening, to exclude studies that clearly met one or more of the exclusion criteria. The remaining articles were then assessed independently by both the first author and a second reviewer for final inclusion. Any differences in judgment were addressed through discussion, and the first author subsequently re-checked the shortlisted studies against the exclusion criteria. In the end, 84 articles were retained and included in the final synthesis stage of the review.

Data analysis

After identifying the final set of eighty-four articles, the authors used a systematic approach to synthesize the literature and extract meaningful insights. First, each reviewer independently read and summarized a subset of the studies. For every article, both reviewers documented key details, including publication year; publication type (e.g., conference paper or journal article) and venue; the engineering sub-discipline(s) and application areas addressed; the first author's country affiliation; any theoretical or conceptual frameworks applied; the study's research focus and methodology; the study setting and system scale; and the main findings. The two sets of summaries were then compared to ensure alignment and consistency across the recorded information.

Next, the authors collaborated to develop a set of inductive codes and a corresponding codebook that captured recurring patterns across the selected studies. The initial codes were informed by dominant topics in the literature, including power grid applications, electric machines, fault diagnosis, real-time implementation, modeling approaches, IoT-based platforms, and cybersecurity. Both reviewers then independently examined the full set of articles again to extract the same summary information as before and to assign each study to one or more of the established codes. When a publication addressed a topic not represented in the existing codebook, additional codes were suggested and incorporated accordingly.

Finally, the codes were examined more closely and clustered into seven broader, emergent themes: power and energy systems applications, electric machines and transportation, fault diagnosis and condition monitoring, real-time control and hardware-in-the-loop (HIL) implementation, modeling and simulation, IoT and computing platforms, and cybersecurity and lifecycle management. An inter-rater reliability metric (e.g., Cohen's Kappa) was applied to evaluate the level of agreement between the two authors when categorizing articles by theme, supporting the consistency and trustworthiness of the coding process. In the final stage of analysis, the first author independently reviewed each publication and assigned it to one or more of the seven emergent themes. The results are reported in two sections. First, descriptive statistics are used to summarize publication trends across the 84 included studies. Second, a qualitative synthesis is presented, organizing and discussing the findings within the seven thematic categories.

Strengths and limitations

This SLR offers a comprehensive overview of DT applications in electrical engineering by examining research trends and the current state of the field. Each emergent theme is paired with implications for both practice and research, providing instructors, researchers, and industry professionals with clear, actionable takeaways. Overall, the findings deepen understanding of this rapidly developing area by highlighting what has been studied to date and outlining promising directions for future work. To the authors' knowledge, no prior SLR has specifically focused on DT research within the broad field of electrical engineering.

Several limitations should be noted. The included studies were selected using exclusion criteria that did not account for quality appraisal or the novelty of findings. In line with similar reviews in engineering education and technical domains (Harris and Kittur, 2024; Kittur et al., 2024; Nguyen and Kittur, 2024), the search was limited to six major databases expected to capture relevant journal and conference publications; however, restricting the search to these sources may have omitted additional relevant work. The review also excluded books, technical reports, and non-English publications, which may have narrowed the scope of evidence captured. Additionally, the selected keywords targeted the overlap between DT technologies and electrical engineering, meaning that studies using alternative terminology may not have been identified. Finally, because the review only included articles published through 2025, some very recent developments may not be reflected in the findings.

Findings

We begin by presenting a descriptive overview of research trends in DT applications within electrical engineering from 2015 to 2025. Next, we detail the seven themes derived from synthesizing the 84 included studies, highlighting representative examples and discussing key implications for both future research and practical implementation.

Descriptive findings related to publication trends

Publications per year

There is a positive trajectory of research interest in DTs within electrical applications over the last decade (see Figure 2). From 2015 to 2025, the number of publications demonstrated an increasing rise in the usage of DT technology in electrical engineering. In the latter half of the 2010s, the number of published articles remained relatively low, but a significant increase is observable beginning in the early 2020s, culminating in a peak output in 2023. This upward trend is highly encouraging, as the rapid acceleration and engagement with DT concepts allow for more integration of DT technology, methodologies, and implementations by researchers in the field. It has become clear throughout our review that DTs have transitioned from a once unknown concept to a core area of applied research. It is likely that this drastic change is due to developments in IoT, cloud computing, and real-time simulation technologies.

Publication type and publication outlet

Of the 84 studies included in this review, most (58.3%) were published as peer-reviewed journal articles, while the remaining 41.7% appeared in conference proceedings. Frequently represented outlets included high-impact, broad-scope journals such as IEEE Access, Energies, Sustainability, and Sensors, along with more specialized venues such as IEEE Transactions on Power Electronics, IEEE Transactions on Industrial Informatics, and IEEE Transactions on Smart Grid. Overall, this spread highlights that DT research is being disseminated through both interdisciplinary engineering journals and discipline-specific electrical engineering publications.

A large portion of the conference papers were published through events sponsored by the Institute of Electrical and Electronics Engineers (IEEE), including the IEEE Power and Energy Society General Meeting (PESGM), the IEEE Applied Power Electronics Conference (APEC), and other IEEE-affiliated international conferences focused on energy systems, transportation, and industrial electronics. The strong representation of IEEE journals and conferences reinforces the electrical engineering orientation of this review and reflects the central role these venues play in reporting emerging applications, case studies, and advancements in DT technologies across the field.

Country affiliation of first author

Table 1 indicates that the studies included in this review had first authors representing 24 different countries. Most first authors were affiliated with China (30.9%), followed by the United States (14.3%), Germany (9.5%), India (7.1%), and Estonia (4.8%). Each of the remaining 23 countries contributed fewer than four publications. It is important to note that the strong representation of DT research from China and the United States may reflect higher levels of research investment and publication volume in these regions across engineering and technology fields. Additionally, researchers and practitioners in other countries may also be actively developing DT solutions, but their work may be shaped by different publication norms, incentives, or levels of institutional and funding support.

Table 1. Sampled articles' country affiliation of the first author.

Theoretical frameworks

We decided to consider the use of theoretical frameworks in DT research for electrical engineering throughout our analysis (please see Table 2). Of these, 38 articles (45.2%) explicitly used a named theoretical framework. The other 46 (54.8%) did not; these focused on technical system implementations, often as technical surveys of methods and technologies. Among the papers employing a formal framework, the multidisciplinary character of the field is reflected in the different methodological approaches. No single framework dominated the literature because none of them appeared in more than three studies. The need for real-time data integration between physical and virtual entities is highlighted in these studies. Nine of the articles without a formal framework were comprehensive reviews or survey papers, which are very common in these kinds of genres.

Table 2. Frameworks used in sampled articles.

Research foci and research methods

An analysis of the primary research objectives and methodological approaches employed across the 84 sampled articles provides a clear picture of the intellectual and practical priorities within the DT research space for electrical engineering. The research foci were categorized inductively based on the central aim or contribution of each publication, while the research methods were classified according to the nature of the data collection and analysis presented. Table 3 summarizes the distribution of these research foci and their associated methodological approaches. Five distinct categories of research focus emerged from the corpus. The largest category, which is System Implementation and Application Case Studies (28.6%), includes articles that primarily describe the development, deployment, and testing of a DT for a specific engineering system, such as a power converter, motor drive, or grid segment. Closely related is the focus on Modeling and Simulation Technique Development (21.4%), where the core contribution is a novel algorithmic, physics-based, or data-driven method for creating or enhancing the DT model itself.

Table 3. Research foci and research methods used in sampled articles.

A significant portion of the literature is dedicated to Framework and Architecture Proposals (22.6%). These articles propose conceptual structures, reference architectures, or integration standards for building and deploying DTs, often discussing interoperability, scalability, and lifecycle management. Another substantial category is Survey and Literature Review (13.1%), which synthesizes existing research, identifies trends, and maps the technological landscape. Finally, a focused subset of articles addresses Fault Diagnosis, Prognostics, and Predictive Maintenance Applications (9.5%), explicitly leveraging the Dt for condition monitoring and reliability engineering. A smaller number of articles (4.8%) explored the development of Educational Tools and Curricula using DT technology.

Regarding research methods, the sample reveals a strong orientation toward empirical and quantitative validation. The most common methodology was Quantitative research (41.7%), employed in studies that presented experimental measurements, simulation results, performance metrics (e.g., efficiency gains, fault detection accuracy), or statistical analyses of system data. A Mixed-Methods approach (28.6%) was also prevalent, particularly in application-focused studies that combined a qualitative description of a proposed framework or system with quantitative data from its implementation or testing.

Qualitative methods (11.9%) were primarily used in review papers, framework proposals, and studies analyzing architectural or conceptual aspects without numerical validation. Studies classified as No Research (Descriptive) (17.9%) were those that provided a detailed technical description of a system, concept, or tutorial without presenting original collected data or formal analysis; this category included some architecture papers and early conceptual explorations. The alignment between focus and method is logical: application and modeling studies heavily favor quantitative validation, framework proposals often use descriptive or mixed methods, and review papers are inherently qualitative. This distribution indicates that while theoretical and architectural discourse is vital, the field is strongly driven by practical implementation and empirical proof-of-concept, underscoring its applied engineering character.

Thematic analysis: descriptions, exemplars, and implications

This section examines in depth the seven emergent themes identified in the review: power and energy systems applications, electric machines and transportation, fault diagnosis and condition monitoring, real- time control and HIL implementation, modeling and simulation, IoT and computing platforms, and cybersecurity and lifecycle management. For each theme, we describe its scope and primary emphasis, clarify how the studies assigned to that theme align with its core ideas, and highlight two representative exemplar papers selected for their clear and direct relevance to the theme. Each thematic section concludes with a brief discussion of key implications for future research and practical application. These exemplar studies were chosen because they place the theme's central concepts at the forefront, unlike other papers that may reference the topic only briefly or indirectly. Table 4 provides definitions for all seven themes, lists the codes associated with each theme (see Appendix A for codes and codes definitions), and reports the number of studies included within each category. A full article-to-theme mapping is included in Appendix B, noting that some studies were assigned to multiple themes when appropriate.

Table 4. Thematic classification of sampled articles.

Figure 3 illustrates how frequently each theme appeared in publications from 2015 to 2025. Notably, no articles published between 2015 and 2017 met the criteria for inclusion in the final review. Among the themes, Power and Energy Systems Applications have consistently received sustained attention across the years. In addition, the results show a growing emphasis on Electric Machines, Drives, and Transportation as well as Real-Time Control, HIL, and Embedded Implementation over time.

Theme 1: power and energy systems applications

Forty-three articles were categorized under the theme of power and energy systems applications, representing the largest and most prominent thematic cluster in this review. This body of work collectively establishes the DT as a transformative tool for the modeling, monitoring, optimization, and management of modern electrical energy infrastructure. The articles address applications across the entire energy value chain, from bulk generation and transmission to distribution networks and distributed energy resources, with a shared goal of improving grid reliability, efficiency, and resilience.

We categorized articles under this theme as describing several key application areas within power systems. A primary focus is on smart grid and utility network operations including studies on power system steady-state modeling (Cui et al., 2020), harmonic analysis for grid construction (Cai et al., 2021), digital power grid definitions (Bai and Wang, 2022), and hierarchical real-time simulation architectures (Ruhe et al., 2023). Research also addresses control center applications (Brosinsky et al., 2018), smart grid dynamics modeling (Teimourzadeh Baboli et al., 2020), and online analysis for power grids (Zhou et al., 2019).

A significant research direction focuses on renewable integration and distributed energy resource (DER) management including fault diagnosis in photovoltaic systems (Jain et al., 2020), grid-tied photovoltaic (PV) system frameworks (Kumar and Kumar, 2025), microgrid system reviews (Kumari et al., 2023), and photovoltaic efficiency enhancement (Hamid et al., 2025). Research also explores energy hub management (Nasiri and Kavousi-Fard, 2023), EV charging station fleets (Francisco et al., 2023), and multi-campus EV grid integration (Erdogan et al., 2025).

The theme includes work on transmission and distribution network applications such as electrical distribution system development (Fernandes et al., 2022), distribution power transformer monitoring (Moutis and Alizadeh-Mousavi, 2021), distribution grid applications (Zhaoyun and Linjun, 2022), and fault location identification (Pamulaparthy, 2024). Research also addresses power quality, stability, and control challenges including impedance-based stability analysis (Lopez Diz et al., 2024), state estimation in inverter-dominated grids (Song et al., 2020), and dynamic state estimation (Idrisov et al., 2025).

Several articles provide comprehensive reviews and framework proposals including foundational explorations (He et al., 2019), power system surveys (Nguyen et al., 2021), future power system overviews (Palensky et al., 2022; Song et al., 2023), electric DT grid technology reviews (Sifat et al., 2023), and smart grid surveys (Xu et al., 2024). Additional applications include power industry case studies (Hu et al., 2020; Huang et al., 2021), power dispatching platforms (Wu et al., 2022), testing grid development (Gonschor et al., 2024), power communication training tools (Guan et al., 2022), broadband signal measurement (Jia et al., 2025), transformer life prediction (Jing et al., 2021), secure transmission schemes (Lian, 2023), digital grid standards (Lin, 2024), IoT applications surveys (Mansour et al., 2023), electric energy metering (Li et al., 2023), smart grid applications (Zhang et al., 2022), UHVDC loss measurement (Yin et al., 2024), monitoring and control services (Kummerow et al., 2020), power system modeling for research (Raskar et al., 2025), data-driven anomaly detection (Shi et al., 2025), and electric power industry reviews (Fu et al., 2022).

Exemplar studies

Exemplar studies under the power and energy systems theme include two articles that highlight the holistic and application-specific value of DTs. One such study is (Cui et al. 2020), who presents a DT framework for power system steady-state modeling, simulation, and analysis. Their work creates a virtual mirror of the physical grid that continuously ingests real-time operational data to perform power flow analysis and “what-if” contingency assessments. This enables grid operators to visualize system status dynamically and test mitigation strategies in a risk-free virtual environment before implementing them physically, exemplifying the core predictive and planning value of the technology.

In another exemplar study, (Jain et al. 2020) focus on a critical reliability challenge: fault diagnosis in distributed PV systems. They develop a data-driven DT approach that compares real-time current-voltage (I-V) characteristics from physical PV arrays with outputs from a detailed, parameterized simulation model. Discrepancies between physical and virtual outputs are used to detect, classify, and locate faults such as short circuits, open circuits, and partial shading. This work powerfully demonstrates the application of DTs for granular asset health monitoring and reliability enhancement in a decentralized, renewable generation context.

Research implications

The extensive work in this theme reveals several critical research frontiers. There is a pressing need for standardized interoperability and data model frameworks (e.g., based on standards like CIM or IEC 61850) to enable DTs from different vendors, or modeling different grid components, to communicate and form a cohesive “system of systems.” Research is also required to overcome computational scalability bottlenecks for system-wide, high-fidelity real-time DTs, potentially through hierarchical, federated, or reduced-order modeling techniques. Furthermore, significant opportunities exist to explore the transition from descriptive/predictive DTs to prescriptive and autonomous control systems capable of initiating self-healing grid actions, requiring advances in real-time control integration and robust decision-making algorithms.

Practice implications

For industry practitioners, system operators, and utility engineers, the studies under this theme validate the DT as an indispensable unified grid intelligence platform. Implementing an operational DT can consolidate disparate data silos (SCADA, GIS, weather, asset management) into a single, coherent virtual model, dramatically enhancing situational awareness. For asset owners and renewable project developers, creating component- level twins for critical equipment like transformers (Jing et al., 2021) or PV inverters enables a shift from costly, calendar-based maintenance to condition-based and predictive maintenance strategies. Furthermore, as shown in (Guan et al. 2022), DTs provide a highly effective, safe, and configurable environment for training new engineers and simulating emergency response procedures, building critical human capacity for modern grid operations.

Theme 2: electric machines, drives, and transportation

Nineteen articles were categorized under the theme of electric machines, drives, and transportation applications. This body of work demonstrates the significant potential of DT technology to enhance the design, control, monitoring, and maintenance of electromechanical energy conversion systems and electrified transportation assets. The articles collectively address how virtual replicas can improve performance, efficiency, reliability, and development cycles for systems ranging from individual motors to complete vehicle propulsion systems.

We categorized articles under this theme as describing several key application areas within electromechanical systems and transportation. The primary focus is on electric vehicle (EV) propulsion and charging infrastructure including studies on propulsion drives for autonomous EVs (Rassolkin et al., 2019), DT platforms for EV applications (Ibrahim et al., 2023), safety system design for EVs (Gokul Krishnan et al., 2023), and EV grid integration frameworks (Erdogan et al., 2025). Research also addresses charging stations for EV fleets (Francisco et al., 2023) and overviews of propulsion drive systems for autonomous vehicles (Ibrahim et al., 2022).

A significant research direction focuses on electric machines and drive systems including DT models of electric drives empowered by Kalman filters (Ebadpour et al., 2023), improved analytical models of induction machines (Mukherjee et al., 2020), electrical motors based on empirical performance models (Rassolkin et al., 2020), and monitoring systems for induction motors using IoT sensors (Santos et al., 2023). Research also explores DTs for electrical machine control and predictive maintenance (Falekas and Karlis, 2021), modeling methods for three-phase inverter-driven motor systems (Song et al., 2024), and service units for AC motor stator fault detection (Rjabtsikov et al., 2021).

The theme includes work on transportation and specialized vehicle applications including electric railway power systems (Ahmadi et al., 2021), light electric vehicle propulsion systems using FPGA implementation (Ruba et al., 2019), complex shipboard DC microgrids (Tavagnutti et al., 2025), all-electric ship power distribution systems (Wong et al., 2023), and fault recovery methods for power electronic converters in transportation applications (Sun et al., 2023). Additional research addresses embedded DTs of power electronic converters for transportation (Wang et al., 2024).

Exemplar studies

Exemplar studies under the electric machines, drives, and transportation theme include two articles that highlight the technical depth and practical implementation challenges of DTs in this domain. One such study is (Ruba et al. 2019), who present a real-time FPGA implementation of a DT for a light electric vehicle propulsion system using Energetic Macroscopic Representation (EMR) organization. Their work demonstrates how DTs can be deployed on embedded hardware for ultra-fast real-time simulation, enabling hardware- in-the-loop testing and potentially onboard diagnostics. This implementation showcases the transition from offline simulation models to embedded, real-time virtual counterparts that can operate synchronously with physical systems.

In another exemplar study, (Falekas and Karlis 2021) provide a comprehensive state-of-the-art review and future prospects for DTs in electrical machine control and predictive maintenance. Their work systematically analyzes how DTs can be leveraged for condition monitoring, remaining useful life prediction, and optimized control of electric machines across various industrial applications. The review identifies key technical challenges including model fidelity, real-time data synchronization, and the integration of machine learning algorithms for adaptive behavior, while outlining a roadmap for future research in this rapidly evolving area.

Research implications

The work in this theme reveals several important research directions. There is a need for high-fidelity, real-time capable models of electric machines and power electronic converters that balance computational efficiency with physical accuracy, particularly for embedded implementation. Research should explore standardized interfaces and data models for electric machine DTs to facilitate interoperability between motor manufacturers, drive suppliers, and system integrators. Significant opportunity exists in developing hybrid modeling approaches that combine physics-based models with data-driven techniques to capture complex phenomena like aging, thermal effects, and fault conditions. Furthermore, research is needed on secure communication protocols for DTs in safety- critical transportation applications, ensuring robust operation in challenging electromagnetic environments.

Practice implications

For manufacturers of electric machines, drives, and vehicle systems, the studies under this theme demonstrate that DTs can significantly accelerate development cycles through virtual prototyping and testing. Engineers can evaluate multiple design iterations, control strategies, and fault scenarios in simulation before physical prototyping, reducing development time and cost. For operators and maintenance teams, implementing DTs enables predictive maintenance strategies for critical assets like traction motors and propulsion drives, potentially preventing costly failures and reducing downtime. In educational contexts, DTs of electric machines and drives can serve as powerful teaching tools, allowing students to interact with virtual representations of complex electromechanical systems, observe internal phenomena that are not visible in physical experiments, and safely explore fault conditions and control responses without risk to equipment.

Theme 3: fault diagnosis, condition monitoring, and maintenance

Fifteen articles were categorized under the theme of fault diagnosis, condition monitoring, and maintenance applications. This body of work demonstrates the significant value of DT technology for enhancing the reliability, availability, and operational efficiency of electrical and electromechanical systems through predictive analytics and intelligent maintenance strategies. The articles collectively address how virtual replicas can enable early fault detection, accurate condition assessment, and optimized maintenance scheduling across diverse engineering assets.

We categorized articles under this theme as describing several key application areas within predictive maintenance and health management. A primary focus is on electrical machine and motor condition monitoring including studies on DTs for electrical machine control and predictive maintenance (Falekas and Karlis, 2021), DT service units for AC motor stator inter-turn short circuit fault detection (Rjabtsikov et al., 2021), and monitoring systems of induction motors using IoT sensors and thermo-magnetic finite element analysis (Santos et al., 2023). Research also addresses improved analytical models of induction machines for DT applications (Mukherjee et al., 2020) and DTs of electrical motors based on empirical performance models (Rassolkin et al., 2020).

A significant research direction focuses on power system and renewable energy asset monitoring including fault diagnosis in distributed photovoltaic systems using DT approaches (Jain et al., 2020), DTs of distribution power transformers for real-time monitoring of medium voltage from low voltage measurements (Moutis and Alizadeh-Mousavi, 2021), and power transformer remaining life prediction based on DT technology (Jing et al., 2021). Research also explores data-driven modeling in DTs for power system anomaly detection (Shi et al., 2025) and high-impedance fault location identification on low voltage distribution systems using IoT-based DT model computation (Pamulaparthy, 2024).

The theme includes work on power electronics and converter health management including controller-embeddable probabilistic real-time DTs for power electronic converter diagnostics (Milton et al., 2020), fault recovery methods for power electronic converters based on accelerator-embedded DTs (Sun et al., 2023), DT-based monitoring and control for DC-DC converters (Lei et al., 2023), and embedded DTs of power electronic converters based on real-time parameter identification (Wang et al., 2024). Additional research addresses mutual information-enhanced DTs for power electronics in smart grid applications (Nalepa and Najdek, 2025) and comprehensive reviews of DT technology applications for electrical equipment lifecycle management (Khalyasmaa et al., 2023).

Exemplar studies

Exemplar studies under the fault diagnosis and condition monitoring theme include two articles that highlight the technical sophistication and practical impact of DTs in predictive maintenance applications. One such study is (Milton et al. 2020), who present controller-embeddable probabilistic real-time DTs for power electronic converter diagnostics. Their work demonstrates how DTs can be implemented directly within converter control hardware to perform continuous, real-time health assessment using probabilistic methods that account for measurement uncertainty and component variability. This approach enables early detection of incipient faults and degradation in critical power conversion equipment, potentially preventing catastrophic failures in applications ranging from renewable energy systems to electric vehicle drivetrains.

In another exemplar study, (Jain et al. 2020) develop a DT approach for fault diagnosis in distributed photovoltaic systems. Their methodology creates detailed simulation models of PV arrays that account for environmental conditions, component characteristics, and system topology, then compares these virtual outputs with real-time measurements from physical systems. Discrepancies between simulated and measured performance are analyzed using machine learning algorithms to identify specific fault types including module failures, string disconnections, and partial shading effects. This work demonstrates how DTs can provide granular, actionable insights for maintenance teams operating large-scale renewable energy installations with thousands of individual components.

Research implications

The research on this theme identifies several important future directions. There is a need for advanced degradation and aging models that can accurately predict remaining useful life of electrical components under varying operational conditions and stress profiles. Research should explore multi-physics coupling approaches that integrate electrical, thermal, mechanical, and environmental factors within DT frameworks to provide comprehensive condition assessment. Significant opportunity exists in developing adaptive and self-learning DTs that continuously update their models based on operational data, improving their accuracy over the asset lifecycle. Furthermore, research is needed on standardized health indicators and prognostic metrics that can facilitate comparison across different asset types and enable portfolio-level maintenance optimization.

Practice implications

For asset owners, operators, and maintenance teams across energy, industrial, and transportation sectors, the studies under this theme demonstrate that DTs enable a fundamental shift from reactive to predictive maintenance paradigms. By continuously comparing actual performance against expected behavior, DTs can identify anomalies long before they develop into functional failures, allowing maintenance to be scheduled during planned outages with minimal disruption. This approach significantly reduces unplanned downtime, extends asset life, and optimizes spare parts inventory. For insurance and risk management professionals, DTs provide quantifiable data on asset health and reliability that can inform premium calculations and coverage terms. In safety-critical applications like electric transportation and power generation, the early fault detection capabilities of DTs contribute to enhanced operational safety by identifying potential failure modes before they compromise system integrity.

Theme 4: real-time control, HIL, and embedded implementation

Fourteen articles were categorized under the theme of real-time control, hardware-in-the- loop (HIL) testing, and embedded implementation of DTs. This body of work demonstrates the critical transition from offline simulation models to real-time executable virtual counterparts that operate synchronously with physical systems, enabling advanced control strategies, hardware validation, and embedded diagnostics. The articles collectively address the technical challenges and implementation architectures for deploying DTs in time-critical applications across electrical engineering domains.

We categorized articles under this theme as describing several key implementation approaches and application areas. A primary focus is on FPGA-based and embedded controller implementations including studies on DT real-time FPGA implementations for light electric vehicle propulsion systems using EMR organization (Ruba et al., 2019), FPGA-based DT implementations for power converter system monitoring (Nwoke et al., 2023), DT-based PID control modeling, design, and development for FPGA implementation (Jain et al., 2024), and embedded DTs of power electronic converters based on GA-PSO real- time parameter identification (Wang et al., 2024). Research also addresses controller-embeddable probabilistic real-time DTs for power electronic converter diagnostics (Milton et al., 2020) and approaches toward embeddable DTs using voxel-based methods for power converter temperature monitoring (Mo et al., 2025).

A significant research direction focuses on hardware-in-the-loop (HIL) and real-time simulation platforms including modular designs and real-time simulators toward power system DTs implementation (Arrano-Vargas and Konstantinou, 2022), design and implementation of hierarchical DTs for power systems using real-time simulation (Ruhe et al., 2023), and DTs of complex shipboard DC microgrids employing the synergy of compiled models and HIL platforms (Tavagnutti et al., 2025). Research also explores DT development of testing grids using grid analyzers (Gonschor et al., 2024) and built-in microcontroller control systems of executive robotic devices using DTs' technology (Vassiliev et al., 2019).

The theme includes work on real-time monitoring and control applications including DT-based monitoring and control for DC-DC converters (Lei et al., 2023), fault recovery methods for power electronic converters based on accelerator-embedded DTs (Sun et al., 2023) and switching converter adaptive control approaches enabled by DTs for all- electric ship power distribution systems (Wong et al., 2023).

Exemplar studies

Exemplar studies under the real-time control and embedded implementation theme include two articles that highlight the technical innovation and practical significance of deploying DTs in time-critical environments. One such study is (Ruba et al. 2019), who present a DT real-time FPGA implementation for light electric vehicle propulsion systems using Energetic Macroscopic Representation (EMR) organization. Their work demonstrates how complex electromechanical system models can be compiled and executed on FPGA hardware with deterministic, microsecond-level latency, enabling true real-time synchronization between physical and virtual systems. This implementation allows for hardware- in-the-loop testing of control algorithms, rapid prototyping of new propulsion configurations, and potential onboard diagnostic applications where computational speed is critical for safety and performance.

In another exemplar study, (Arrano-Vargas and Konstantinou 2022) explore modular design and real-time simulators toward power system DTs implementation. Their research addresses the significant challenge of scaling DTs to large, complex power systems while maintaining real-time execution capability. By employing modular architectures and optimized simulation algorithms, they demonstrate how system-level DTs can operate with sufficient speed to support online decision-making and control applications. This work is particularly significant for smart grid applications where dynamic phenomena like transient stability and power quality events require simulation timescales that match or exceed real-time operation for effective monitoring and control.

Research implications

The research on this theme identifies several critical future directions. There is a pressing need for standardized interfaces and communication protocols that enable seamless integration between DTs and various real-time control hardware platforms (FPGAs, microcontrollers, PLCs, RTUs). Research should explore model reduction and computational optimization techniques that preserve essential system dynamics while enabling faster-than-real-time simulation for predictive control applications. Significant opportunity exists in developing adaptive real-time synchronization algorithms that can maintain alignment between physical and virtual systems despite communication delays, measurement noise, and system perturbations. Furthermore, research is needed on cyber-physical security frameworks for embedded DTs operating in safety-critical applications, ensuring protection against both cyber threats and potential model-based attacks.

Practice implications

For control system engineers, power electronics designers, and embedded systems developers, the studies under this theme demonstrate that DTs enable accelerated development and validation cycles through hardware-in-the-loop testing. Engineers can evaluate control algorithms, fault response strategies, and system interactions in a virtual environment before deployment on physical hardware, reducing development risk and time-to-market. For system operators and maintenance teams, embedded DTs provide continuous, non-intrusive monitoring capabilities that operate alongside control systems without interfering with primary functions. In safety-critical applications like electric transportation and grid protection systems, real-time DTs enable predictive control and fault anticipation strategies that can prevent catastrophic failures by identifying abnormal conditions before they escalate. The implementation of DTs on embedded platforms also facilitates edge computing applications where processing occurs locally rather than in the cloud, reducing latency, bandwidth requirements, and vulnerability to communication disruptions.

Theme 5: modeling, simulation, and system identification

Thirteen articles were categorized under the theme of modeling, simulation, and system identification approaches for DTs. This body of work represents the foundational technical core of DT technology, focusing on the methods, algorithms, and frameworks used to create accurate virtual representations of physical systems. The articles collectively address the fundamental challenge of developing mathematical models that faithfully capture system behavior across various operating conditions and time scales.

We categorized articles under this theme as describing several key methodological approaches to DT model development. A primary focus is on physics-based and analytical modeling techniques including studies on improved analytical models of induction machines for DT applications (Mukherjee et al., 2020), DT models of power electronic converters (Sado et al., 2023), and DT models of power electronic converters using dynamic neural networks (Wunderlich and Santi, 2021). Research also addresses the DT of discrete dynamic systems, examining initial approaches and future challenges (Ganguli and Adhikari, 2020), and DT modeling methods of three-phase inverter-driven PMSM systems for parameter estimation (Song et al., 2024).

A significant research direction focuses on data-driven and hybrid modeling approaches including mutual information-enhanced NARX-NN DTs for power electronics in smart grid applications (Nalepa and Najdek, 2025), DT frameworks for grid- tied PV systems using frequency domain modeling and E2FD-HO-based approaches for power electronic circuits (Kumar and Kumar, 2025), and DT-based frameworks for load identification using odd harmonic current plots (Mylona and Bouhouras, 2025). Research also explores practical approaches to construct DTs of power grids using harmonic spectra (Cai et al., 2021) and preliminary explorations of DTs for power systems addressing challenges, frameworks, and applications (He et al., 2019).

The theme includes work on system identification and parameter estimation methods including DTs of electrical motors based on empirical performance models (Rassolkin et al., 2020), dynamic state estimation in power electronics-dominated grids using DT approaches (Idrisov et al., 2025), and comprehensive surveys of DT techniques for power electronics-based energy conversion systems examining concepts, application scenarios, and future trends (Chen et al., 2022). Additional research addresses visions toward standard architectures for future power system DTs (Zomerdijk et al., 2024).

Exemplar studies

Exemplar studies under the modeling and simulation theme include two articles that highlight contrasting yet complementary approaches to DT model development. One such study is (Ganguli and Adhikari 2020), who provide a fundamental examination of DTs for discrete dynamic systems, analyzing initial approaches and future challenges in physics-based modeling. Their work systematically addresses the mathematical foundations of creating DTs for systems with discontinuous or switching behavior—common in power electronics and digital control systems—and identifies key challenges in maintaining model fidelity while ensuring computational tractability. This research provides important theoretical grounding for the development of accurate simulation models that can capture both continuous dynamics and discrete events.

In another exemplar study, (Nalepa and Najdek 2025) present mutual information- enhanced NARX-NN (Nonlinear AutoRegressive with eXogenous inputs Neural Network) DTs for power electronics in smart grid applications. Their work demonstrates a sophisticated data-driven approach where neural network architectures are enhanced with mutual information criteria to select the most relevant historical inputs for predicting system behavior. This methodology represents a significant advancement in creating adaptive, learning-based DTs that can capture complex nonlinear relationships in power electronic systems without requiring complete first-principles knowledge of all underlying physics. The approach is particularly valuable for systems where certain parameters or behaviors are difficult to model analytically or vary unpredictably over time.

Research implications

There is a pressing need for systematic model validation and verification frameworks that can quantify the accuracy and uncertainty of DT predictions across different operating regimes. Research should explore multi-fidelity modeling approaches that combine detailed physics-based models for critical components with faster data-driven surrogates for less critical elements, enabling system-level simulation with manageable computational demands. Significant opportunity exists in developing adaptive parameter estimation algorithms that can continuously update model parameters based on operational data, compensating for component aging, environmental changes, and other time-varying effects. Furthermore, research is needed on standardized model exchange formats and interfaces that facilitate reuse and interoperability of DT components across different simulation platforms and application domains.

Practice implications

For modeling and simulation specialists, control system designers, and research engineers, the studies under this theme demonstrate that effective DT implementation requires careful balancing of model complexity, accuracy, and computational requirements. The choice between physics-based, data-driven, or hybrid approaches should be guided by the specific application requirements, available data, and computational constraints. For system integrators and solution architects, understanding different modeling paradigms enables informed selection of appropriate techniques for different subsystems within a larger DT implementation. In educational and training contexts, the methodological diversity evident in this theme provides rich pedagogical material for teaching modeling principles, demonstrating how different approaches, from first-principles physics to machine learning, can be applied to solve similar engineering challenges. The development of standardized modeling approaches also facilitates knowledge transfer and collaboration across organizations and research groups working on similar DT applications.

Theme 6: IoT, data, and computing platforms

Twelve articles were categorized under the theme of Internet of Things (IoT), data management, and computing platforms for DTs. This body of work addresses the critical infrastructure layer that enables DT functionality—the sensors, networks, data pipelines, and computational resources that connect physical assets to their virtual counterparts. The articles collectively examine how modern computing architectures and data technologies support the creation, operation, and scaling of DT systems across electrical engineering applications.

We categorized articles under this theme as describing several key technological enablers for DT implementations. A primary focus is on IoT sensing and connectivity infrastructures including studies on built-in microcontroller control systems of executive robotic devices using DTs technology (Vassiliev et al., 2019), DT-based monitoring systems of induction motors using IoT sensors and thermo-magnetic finite element analysis (Santos et al., 2023), high-impedance fault location identification on low voltage distribution systems using IoT-based DT model computation (Pamulaparthy, 2024), and automated behavior analysis methods for large-scale distributed IoT systems using DT (Sleuters et al., 2019). Research also explores DT frameworks for electric vehicle grid integration through multi-campus case studies integrating smart wireless and renewable energy systems (Erdogan et al., 2025).

A significant research direction focuses on edge, cloud, and distributed computing platforms including integration of 6G signal processing, communication, and computing based on information timeliness-aware DTs (Liao et al., 2024), DT-6G empowered future smart grid applications (Kharbouch et al., 2025), and approaches toward embeddable DTs using voxel-based methods for power converter temperature monitoring (Mo et al., 2025). Research also examines approaches integrating model-based systems engineering, IoT, and DTs for the design of electric unmanned autonomous vehicles (Ramirez et al., 2025) and applications of IoT and DTs in electrical power systems through comprehensive surveys (Mansour et al., 2023).

The theme includes work on specialized computing and data processing applications including DT assistant-systems for state estimation in inverter dominated grids (Song et al., 2020), advancing electrical engineering education with digital tools for electromagnetic characterization (Bojita et al., 2025), and DTs of electrical stimulation devices to monitor and control electrical stimulation of cells in vitro (Zimmermann et al., 2021).

Exemplar studies

Exemplar studies under the IoT and computing platforms theme include two articles that highlight the integration of cutting-edge computing technologies with DT implementations. One such study is (Liao et al. 2024), who explore the integration of 6G signal processing, communication, and computing based on information timeliness-aware DTs. Their research addresses the critical challenge of latency in cyber-physical systems by leveraging emerging 6G network capabilities to ensure that DTs receive sensor data with minimal delay and that computational results are returned to physical systems within required time constraints. This work is particularly significant for real-time control applications where synchronization between physical and virtual systems depends on ultra-reliable, low-latency communication links.

In another exemplar study, (Sleuters et al. 2019) present a DT method for automated behavior analysis of large-scale distributed IoT systems. Their approach demonstrates how DTs can serve as computational frameworks for analyzing complex interactions within massive IoT deployments, identifying normal operational patterns, and detecting anomalies across distributed sensor networks. This work highlights the data management and processing challenges inherent in scaling DTs to large systems with thousands of interconnected components, and proposes algorithmic approaches for extracting meaningful insights from high-volume, high-velocity IoT data streams while maintaining computational efficiency.

Research implications

There is a need for adaptive data ingestion and preprocessing frameworks that can handle heterogeneous data streams from diverse sensor types with varying sampling rates, formats, and quality levels. Research should explore edge-cloud computing architectures that optimally distribute computational tasks between local edge devices and centralized cloud resources based on latency requirements, data privacy considerations, and computational complexity. Significant opportunity exists in developing semantic data models and ontologies that enable interoperability between DTs, IoT platforms, and legacy systems across different vendors and application domains. Furthermore, research is needed on scalable data storage and retrieval systems that can efficiently manage the historical data required for training data-driven models while supporting fast querying for real-time decision support.

Practice implications

For system architects, data engineers, and IoT solution providers, the studies under this theme demonstrate that successful DT implementation requires robust data infrastructure that extends beyond the simulation models themselves. The choice of sensors, communication protocols, data storage solutions, and computing platforms fundamentally determines the performance, scalability, and maintainability of DT systems. For organizations implementing DTs, establishing data governance and quality assurance processes is critical for ensuring that models receive accurate, timely, and consistent data from physical assets. In educational settings, this theme highlights the interdisciplinary nature of DT implementation, requiring integration of knowledge from sensing technologies, networking, data science, and domain-specific engineering disciplines. The emphasis on computing platforms also points to opportunities for standardized middleware and platform-as-a- service offerings that can reduce implementation complexity and accelerate adoption across different application areas within electrical engineering.

Theme 7: cybersecurity, reliability, and lifecycle management

Five articles were categorized under the theme of cybersecurity, reliability, and lifecycle management frameworks for DT-enabled systems. This body of work addresses the higher-level, systemic considerations that extend beyond individual technical implementations to encompass security, dependability, standardization, and comprehensive asset management throughout the system lifecycle. The articles collectively examine how DTs can enhance—and must be designed to ensure—the security, resilience, and sustainable management of critical engineering infrastructure.

We categorized articles under this theme as describing several key systemic frameworks and considerations for DTs. A primary focus is on cybersecurity and secure system architectures including studies on secure transmission schemes for smart grid cyber-physical systems with DT architectures based on multi-agent deep reinforcement learning (Lian, 2023) and DT-6G empowered future smart grid applications examining the intersection of advanced networking and security (Kharbouch et al., 2025). These works address the dual role of DTs as both potential security vulnerabilities requiring protection and as tools for enhancing system security through simulation of cyber-attack scenarios and defense strategies.

A significant research direction focuses on standardization, taxonomy, and architectural frameworks including visions toward standard architectures for future power system DTs (Zomerdijk et al., 2024) and construction of new digital grid standard systems for power systems based on DT technology (Lin, 2024). These articles propose systematic approaches to DT design, implementation, and interoperability, addressing the need for common reference models, data formats, and interface specifications that can facilitate widespread adoption and integration across different systems and organizations.

The theme includes work on reliability, resilience, and comprehensive lifecycle management including reviews of DT technology applications for electrical equipment lifecycle management (Khalyasmaa et al., 2023). This comprehensive examination addresses how DTs can support asset management throughout all phases of the lifecycle—from design and commissioning through operation and maintenance to decommissioning and replacement—by providing continuous virtual representations that evolve alongside physical assets.

Exemplar studies

Exemplar studies under the cybersecurity and lifecycle management theme include two articles that highlight the strategic importance of addressing security and standardization in DT implementations. One such study is (Lian 2023), who presents a secure transmission scheme for smart grid cyber-physical systems with DT architecture based on multi-agent deep reinforcement learning. This research uniquely positions the DT not merely as a monitoring tool but as an active component of the cybersecurity defense system. By creating a virtual replica of the physical grid that includes network communication models, the system can simulate various cyber-attack scenarios in real-time, train reinforcement learning agents to recognize attack patterns, and develop optimal defense strategies that can be deployed to protect the actual physical infrastructure. This approach demonstrates how DTs can transform cybersecurity from reactive to proactive discipline.

In another exemplar study, (Zomerdijk et al. 2024) present a vision toward standard architecture for future power system DTs. Their work addresses a critical barrier to DT adoption: the proliferation of proprietary, incompatible implementations that hinder interoperability and data exchange between systems from different vendors or organizations. By proposing a standardized architectural framework, they provide a roadmap for creating DTs that can seamlessly integrate with existing utility systems, share information across organizational boundaries, and scale to address increasingly complex grid management challenges. This standardization effort is particularly important for critical infrastructure where long-term interoperability, maintainability, and evolution of systems are essential considerations.

Research implications

There is a pressing need for comprehensive security frameworks that address not only traditional IT security concerns but also the unique vulnerabilities introduced by the tight coupling of physical and virtual systems in DT architectures. Research should explore quantitative metrics for DT reliability and resilience that can inform design decisions and operational risk assessments. Significant opportunities exist in developing lifecycle management methodologies that leverage DTs to optimize total cost of ownership, including energy efficiency, maintenance scheduling, and end-of-life decisions. Furthermore, research is needed on regulatory and policy frameworks that address liability, data ownership, and certification requirements for DTs in safety-critical applications like power systems and transportation infrastructure.

Practice implications

For system architects, security specialists, and asset managers, the studies under this theme demonstrate that DTs must be designed with security and lifecycle considerations as foundational requirements rather than afterthoughts. Implementing security-by-design principles and establishing clear data governance protocols are essential for protecting critical infrastructure from both cyber and physical threats. For standards organizations and regulatory bodies, this research highlights the importance of developing interoperability standards and certification processes that can ensure DT implementations meet minimum requirements for security, reliability, and data quality. In organizational contexts, adopting comprehensive lifecycle management approaches enabled by DTs can transform asset management from a cost center to a value-generating activity by optimizing maintenance expenditures, extending asset life, and improving operational efficiency. The emphasis on standardization also suggests opportunities for creating reference implementations and best practice guidelines that can accelerate adoption while ensuring consistent quality and security across different implementations within an organization or industry sector.

Discussion

While the seven themes presented in the Findings section are distinct, several papers in this review naturally examine intersections across themes, demonstrating the multifaceted nature of DT applications in electrical engineering. For example, (Jain et al. 2020) examined the intersection of power and energy systems (Theme 1) and fault diagnosis (Theme 3) while developing a DT for fault detection in distributed photovoltaic systems. In another study, (Ruba et al. 2019) examined the integration of electric machine modeling (Theme 2) with real-time embedded implementation (Theme 4) while implementing an FPGA-based DT for a light electric vehicle propulsion system, representing work at the intersection of transportation applications and hardware implementation. Finally, Lian's (2023) work on secure transmission schemes for smart grid cyber-physical systems with DT architecture lay at the intersection of power systems (Theme 1), cybersecurity (Theme 7), and advanced computing platforms (Theme 6), as the main goals of the study were understanding how to secure DT-enabled infrastructure while leveraging machine learning for enhanced protection.

Further examination of the intersections of the themes identified in this SLR generates additional observations and implications for research and practice. For example, a potential research question emerging from the integration of Themes 1, 5, and 7 could be: what standardized modeling frameworks (Theme 5) for power system DTs (Theme 1) most effectively support cybersecurity assessments and resilience planning (Theme 7) while maintaining real-time performance? Studies could also be conducted to better understand the relationship between high-fidelity modeling techniques (Theme 5) and their implementation requirements (Theme 4) as they relate to the accuracy and latency of fault diagnosis systems (Theme 3) for safety-critical applications like electric transportation (Theme 2). Finally, further examining the data infrastructure requirements (Theme 6) for scaling DTs from component-level to system-wide implementations across different engineering domains represents another promising direction for research motivated by the intersection of identified themes.

Despite the diversity of application domains represented by the seven themes, a common challenge emerges across all categories: the trade-off between model fidelity and real-time computational performance. In power and energy systems (Theme 1), high-fidelity electromagnetic transient models provide accurate results but cannot yet run in real-time for system-wide grids. In electric machines and transportation (Theme 2), detailed finite element models capture thermal and magnetic dynamics but are too slow for onboard embedded implementation. Fault diagnosis systems (Theme 3) require both accuracy and speed to detect incipient failures before they escalate, yet high-fidelity models introduce latency that delays detection. Real-time control and HIL implementations (Theme 4) confront this trade-off most directly, often sacrificing physical accuracy to achieve microsecond-level execution. Modeling approaches (Theme 5) attempt to resolve this tension through reduced-order models and surrogate techniques, but no universally accepted method exists. IoT and computing platforms (Theme 6) push computation to the edge or cloud to balance these demands, yet communication latency introduces new challenges. Cybersecurity frameworks (Theme 7) must model both physical and cyber domains simultaneously, compounding the fidelity-speed dilemma. This shared tension across all seven themes suggests that future research should prioritize adaptive multi-fidelity architectures that dynamically adjust model complexity based on operational context and real-time requirements, rather than pursuing one-size-fits-all solutions.

Our descriptive analysis of the articles also highlighted significant opportunities for methodological advancement and broader impact. The diverse range of theoretical and conceptual frameworks identified in Table 2, from physics-based finite element analysis (Santos et al., 2023) to data-driven neural networks (Wunderlich and Santi, 2021) and standardized architectural models (Zomerdijk et al., 2024), reveal a vibrant but fragmented methodological landscape. This presents an opportunity to develop more integrated, hybrid frameworks that combine the strengths of different approaches while establishing common validation protocols.

Additionally, studies focused on fundamental modeling research (Theme 5) were more likely than those focused on specific applications to employ advanced methodological techniques and larger sample sizes or validation datasets. As the DT research community matures, these trends suggest an opportunity to elevate application studies through more rigorous validation methodologies and cross-institutional collaboration to improve generalizability.

The geographical distribution of research output, with strong contributions from China (30.9%), the United States (14.3%), and Germany (9.5%), indicates both concentrated expertise and potential gaps in global representation. This distribution may influence the specific application contexts being studied and the regulatory or standardization frameworks being considered. Furthermore, only a limited number of studies from the sampled articles explicitly focused on educational applications or workforce development considerations (Bojita et al., 2025; Guan et al., 2022); this indicates a need for more research on how DT technology can be effectively integrated into engineering curricula to prepare the next generation of engineers for increasingly virtualized design and operation environments.

This SLR has several limitations that readers should consider when interpreting its findings. First, the selection of the final 84 articles was based on exclusion criteria that did not include a formal metric of quality or uniqueness of information contained within them. Consequently, all included studies were treated with equal weight regardless of methodological rigor, sample size, or validation depth. Second, while six reputable databases were used, including Google Scholar, Web of Science, IEEE Xplore, ERIC, Compendex, ScienceDirect, and Wiley Online Library, books, technical reports, and non-English publications were excluded, which may have limited the scope of information covered within the SLR. Third, the search terms focused specifically on the intersection of DT technology with electrical engineering and related subfields such as power systems, smart grids, and embedded systems; other relevant articles might have emerged using different keyword combinations, such as cyber-physical system or virtual commissioning. Fourth, because articles published after 2025 were not included in the systematic review, the review may be missing recent developments that practitioners and researchers might find relevant but were not yet published. Fifth, the high occurrence of DT research originating from China (30.9%) and the United States (14.3%) in our search could be influenced by broader research investment and publication output in these regions, potentially underrepresenting work from countries with different publication practices, incentives, or research support infrastructure. Sixth, as with any systematic literature review, the coding and thematic categorization involved subjective judgments by the authors; although inter-rater reliability measured by Cohen's Kappa was calculated to confirm dependability, alternative categorizations are possible. Seventh, the broad scope of electrical engineering as a field means that this review necessarily scratches the surface of each thematic area; deeper, domain-specific systematic reviews focused exclusively on power electronics or electric machines, for example, would be valuable complementary studies.

Conclusion

This systematic literature review set out to answer the overarching research question stated in the abstract: what is the current state of research on DT applications in electrical engineering? Based on the synthesis of 84 articles published between 2015 and 2025, the current state is characterized by five key insights. First, power and energy systems represent the most mature and heavily populated application domain, with 43 articles demonstrating DT utility for grid monitoring, renewable integration, and stability analysis. Second, Internet of Things infrastructure and computing platforms serve as critical enabling foundations, yet remain fragmented in implementation. Third, the field balances physics-based and data-driven modeling approaches, though hybrid methods are emerging as a promising direction. Fourth, real-time and embedded implementations for control and diagnostics are gaining traction but face scalability and latency barriers. Fifth, cybersecurity, standardization, and lifecycle management, while underrepresented with only five articles, are increasingly recognized as essential for long-term viability. Regarding future integration and scalability, this review identifies that standardized interoperability frameworks such as CIM and IEC 61850, hierarchical and federated twin architectures, and adaptive real-time synchronization algorithms are the most pressing enablers for moving from component-level demonstrations to system-wide, scalable deployments. Thus, this review successfully maps the current knowledge landscape and pinpoints specific, actionable directions for future research and practice, directly fulfilling the stated objective.

In summary, this review makes three core contributions to electrical engineering research. First, it provides the first dedicated systematic literature review on DTs in electrical engineering, revealing that while 84 studies exist across seven thematic areas, only five percent address educational applications and only six percent focus on cybersecurity, indicating critical underinvestment. Second, it establishes that the transition from descriptive to prescriptive DTs remains the central technical barrier, requiring breakthroughs in real-time control integration and decision-making algorithms. Third, it demonstrates that without standardized architectures and validated interoperability protocols, DT adoption will remain siloed and non-scalable. For practitioners, the message is clear: DTs are no longer conceptual but rather operational tools for predictive maintenance, hardware-in-the-loop validation, and grid resilience, but only when implemented with rigorous attention to data infrastructure, model fidelity, and security-by-design principles. For researchers, the roadmap is equally clear: prioritizing hybrid modeling, federated architectures, and lifecycle management frameworks. Without these advances, DT technology risks remaining a collection of promising prototypes rather than a transformed engineering paradigm.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this work, the author(s) used ChatGPT 5o to improve the language and readability. After using this tool/service, the author(s) reviewed and edited the content as needed and take full responsibility for the content of the publication.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frsgr.2026.1805035/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Athula Rajapakse, University of Manitoba, Canada

Reviewed by: Mihiran Galagedarage Don, Memorial University of Newfoundland, Canada

Adrian Gligor, George Emil Palade University of Medicine, Pharmacy, Sciences and Technology of Târgu Mureş, Romania