Artificial Intelligence in Smart Grids — A Comprehensive Survey

Rachmad Andri Atmoko
Head of Laboratory Internet of Things and Human Centered Design
Universitas Brawijaya, Indonesia
[e-mail]

The transformation of electricity grids into smart grids is a critical evolution in modern energy systems. Artificial Intelligence (AI) has emerged as a pivotal enabler of this transformation, offering advanced methodologies for automation, optimization, and decision-making. AI applications in smart grids span from predictive maintenance, demand response optimization, renewable energy integration, real-time data processing, to grid flexibility. This chapter provides a holistic review of the current state of AI applications in smart grids, analyzing the latest research, methodologies, and practical implementations extracted from a deep literature survey.

AI Applications in Smart Grids

Predictive Maintenance

Predictive maintenance has become a transformative approach in enhancing the efficiency and reliability of grid infrastructure by utilizing AI-driven technologies. By leveraging advanced machine learning algorithms, particularly deep learning and anomaly detection techniques, AI systems can assess the health of grid infrastructure in a more proactive and effective manner. Studies show that AI methods significantly outperform traditional failure detection systems, which often rely on basic threshold-based rules and manual monitoring. With the power of deep learning, AI models can detect subtle anomalies and patterns that might otherwise go unnoticed, enabling early identification of potential failures before they escalate into costly outages [3,14].

Additionally, Deep Reinforcement Learning (DRL) has proven to be a game-changer in predictive maintenance applications, allowing for real-time fault detection and dynamic decision-making. DRL can adapt continuously to changing conditions, improving the accuracy of fault identification and reducing response time, which is critical for maintaining grid stability and minimizing downtime [13]. This approach enhances the capability of predictive maintenance systems to detect and respond to emerging issues in real-time, thus optimizing maintenance schedules and resource allocation.

Furthermore, hybrid neural network models are emerging as a promising solution to tackle the complex nature of grid systems. These models combine multiple neural network architectures to enhance feature learning, enabling them to capture multiple, simultaneous faults that may occur across the grid more effectively than traditional threshold-based detection models. This approach allows for a more nuanced understanding of the system’s operational state, identifying not just single faults but interrelated issues that could impact the grid’s performance. This integrated approach significantly improves predictive maintenance outcomes, providing a more comprehensive view of system health [12].

However, the widespread adoption of AI for predictive maintenance in grid infrastructure faces significant challenges. One of the primary obstacles is the need for real-time data integration across various systems. The lack of standardization and interoperability among different data sources and equipment makes it difficult to fully integrate AI models into existing infrastructure. As a result, the ability to deploy AI solutions effectively is often limited by these technical barriers, hindering the potential of predictive maintenance to scale across industries and regions. Addressing these issues requires collaboration between industry stakeholders to establish standardized data protocols and ensure seamless communication between systems.

Demand Response Optimization

Demand response (DR) strategies have emerged as a critical component in modernizing the energy grid, allowing for more flexible operations by adjusting energy consumption patterns based on grid constraints and economic signals. These strategies are designed to balance supply and demand efficiently while enhancing grid stability and reducing operational costs. AI-based DR management is particularly valuable, as it employs advanced prediction models and adaptive control strategies to optimize energy consumption. For instance, machine learning models like Convolutional Neural Networks combined with Long Short-Term Memory (CNN-LSTM) networks are able to accurately forecast consumption patterns by taking into account various factors, including consumer behavior and energy pricing. These models improve the precision of load predictions, enabling more effective demand-side management [7].

In addition to predictive models, heuristic optimization methods, such as Genetic Algorithms (GA) and Particle Swarm Optimization (PSO), are employed to improve the scheduling of loads. These optimization techniques enhance cost efficiency by determining the most economical way to distribute energy use across time, particularly in response to changing demand and pricing signals. By doing so, they help reduce costs associated with energy consumption while supporting a more flexible and dynamic grid [8]. Moreover, Multi-agent Systems (MAS) are increasingly being used to coordinate Distributed Energy Resources (DER) for dynamic load balancing. These systems facilitate the integration and efficient management of multiple energy sources, improving the overall coordination and operation of the grid [1].

Key findings from recent studies indicate that AI-based DR strategies play a significant role in reducing peak loads and operational costs by enhancing demand-side flexibility. This is achieved through improved prediction and adaptive control mechanisms that better align energy consumption with grid needs, thereby mitigating the risks associated with overloading and inefficient energy use. Additionally, the adoption of dynamic energy pricing strategies supported by AI can further optimize energy consumption patterns, promoting more sustainable and cost-effective energy use. AI-based recommendations are also able to provide personalized incentives, encouraging consumers to adjust their energy consumption behavior in response to both grid conditions and economic signals, thus further improving overall efficiency and reducing unnecessary demand spikes [15,9].

However, despite these advancements, there are several challenges that hinder the scalability of AI-driven demand response strategies. One of the primary concerns is the issue of data privacy, particularly when dealing with real-world data that includes detailed consumer usage patterns. Ensuring the protection of sensitive data while enabling effective demand-side management is crucial for widespread adoption. Additionally, there are constraints related to grid-scale deployment, such as the integration of AI solutions into existing infrastructure and the coordination of numerous energy sources and agents across a large-scale network. These challenges highlight the need for continued innovation in AI technologies and regulatory frameworks that balance efficiency with privacy and security concerns [2,4].

Renewable Energy Integration

Renewable energy integration into the smart grid presents significant challenges due to the inherent variability and intermittency of energy sources such as wind and solar power. These sources are not only unpredictable but can also cause fluctuations in energy generation, making it difficult to maintain a stable and reliable grid. However, AI technologies have made substantial contributions in addressing these challenges by enabling robust forecasting and balancing strategies that enhance grid performance and sustainability. One such innovation involves hybrid deep learning models, such as Convolutional Neural Networks combined with Autoencoders and Long Short-Term Memory (CNN-Autoencoder-LSTM) networks. These models have been shown to improve solar and wind energy forecasting accuracy by up to 30%, outperforming traditional standalone models. By leveraging these hybrid models, AI systems can predict renewable energy output more effectively, helping grid operators anticipate fluctuations and adjust energy distribution accordingly [14].

In addition to forecasting improvements, Physics-Informed Neural Networks (PINNs) are another breakthrough in AI-driven energy management. PINNs incorporate real-world grid constraints into AI models, enabling more accurate energy scheduling and distribution by aligning predictions with physical limitations of the grid infrastructure. This approach not only reduces errors in energy management but also ensures that renewable energy is optimally integrated into the grid, minimizing wasted potential [5]. Moreover, AI technologies play a crucial role in optimizing Optimal Power Flow (OPF), which is a critical function in balancing energy generation and consumption. AI-based OPF algorithms can respond to fluctuations in renewable energy output almost instantaneously, ensuring that the grid remains stable even during periods of high variability in renewable generation. This capability is essential for maintaining a reliable and efficient grid while supporting the continued growth of renewable energy adoption [11].

The outcomes of these AI-driven solutions are significant in advancing renewable energy integration into smart grids. One of the key outcomes is the enhanced minimization of curtailment, where surplus renewable energy that would typically be wasted is better predicted and allocated. By more accurately forecasting renewable energy production, AI can ensure that excess energy is either stored or redirected to where it is needed most, thereby reducing energy waste and improving overall system efficiency [5]. Additionally, improved forecasting models contribute to better balancing strategies for grids that rely heavily on renewable energy sources, increasing grid stability and enhancing long-term sustainability. These advancements not only facilitate the integration of renewables but also support a more resilient and sustainable energy infrastructure [7,18].

However, despite the progress made, there remain significant challenges to the widespread scalability of AI models in grid management. While most forecasting solutions perform well at a local level, their application to large-scale, grid-wide systems presents substantial computational hurdles. The need for real-time data processing and the complexity of managing diverse energy sources across a broad geographical area require significant computational power. This remains a key concern for deploying AI-based solutions on a larger scale, as grid operators must ensure that these systems can handle the vast amounts of data and the computational demands required for effective energy management across entire networks [12].

Real-Time Data Processing

The increasing number of Internet of Things (IoT) devices, smart meters, and sensors integrated into modern electricity networks has made real-time data processing a critical component in the effective operation of smart grids. With these devices continuously collecting vast amounts of data, it is essential to have advanced AI-driven systems in place to process, analyze, and act on this information in real time. AI plays a significant role in enabling more efficient grid management by leveraging cutting-edge techniques such as federated learning, deep learning models, and transformer-based forecasting. Federated learning, for instance, allows for the distributed training of AI models across edge nodes in the network, ensuring that data can be processed locally without the need to share raw data between devices or centralized systems. This method not only enhances the speed and efficiency of data processing but also preserves the privacy of sensitive consumer data, which is a critical concern in smart grid applications [12].

In addition to federated learning, deep learning models are employed for large-scale energy consumption analytics and anomaly detection. These models are designed to handle vast quantities of data and are capable of identifying patterns and irregularities in energy use that may indicate potential issues, such as equipment failures or energy theft. By leveraging deep learning, smart grids can detect anomalies with greater speed and accuracy, enabling grid operators to take corrective actions before minor issues escalate into major disruptions [3]. Furthermore, transformer-based forecasting techniques are increasingly used to analyze real-time grid data, as they are particularly effective at extracting time-dependent patterns from massive datasets. These models allow for better prediction of energy demand and generation, helping to optimize grid operations by accurately forecasting fluctuations in energy supply and consumption [7].

Despite these advancements, several challenges remain in the development and deployment of AI models for smart grids. One of the primary challenges is the limited research on scalable AI models that can efficiently process decentralized, high-velocity grid data while meeting latency constraints. The decentralized nature of smart grids, with data originating from various IoT devices and sensors, requires AI models that can handle the complexity of such data distribution without compromising speed or accuracy. Moreover, the real-time processing demands of these systems make it difficult to balance the computational resources required for efficient analysis with the need for low-latency responses, especially in large-scale grid networks. As a result, further research is needed to develop AI solutions that can effectively process this high-velocity data while ensuring that performance remains reliable and scalable across a broad, distributed grid infrastructure [12].

*Grid Flexibility and Adaptive Control
*
AI plays a crucial role in enabling dynamic grid operations, facilitating distributed energy trading, and optimizing grid flexibility parameters in near real-time. One of the key AI technologies enhancing grid performance is Multi-Agent Reinforcement Learning (MARL). This approach optimizes the coordination of distributed energy resources, battery scheduling, and adaptive microgrid control. By utilizing MARL, energy systems can dynamically adjust to changing conditions, ensuring more efficient energy distribution and reducing costs associated with energy storage and grid management [1,5]. This adaptive control mechanism is particularly important in decentralized energy systems, where numerous small-scale generators and storage units must work together to meet demand and maintain grid stability.

Another significant AI-driven advancement in grid operations is AI-based Dynamic Line Rating (DLR). DLR technologies enable real-time adjustments to grid transmission capacity by taking into account current weather conditions and real-time load factors. These adjustments allow for enhanced operational flexibility, enabling grids to handle greater fluctuations in demand without overloading the system. By using real-time data on temperature, wind speed, and other environmental factors, DLR systems can optimize the flow of electricity through the grid, reducing the risk of congestion and enhancing overall grid reliability [5].

One of the most exciting developments in AI for smart grids is the emergence of energy flexibility trading platforms. These platforms, powered by AI, allow consumers to bid on demand flexibility based on probabilistic AI models. Through these platforms, consumers can adjust their energy consumption patterns in response to grid needs, selling excess flexibility back to the grid or receiving incentives for reducing their demand during peak periods. This type of dynamic energy trading has been tested in real-world trials, where AI models predict the optimal times for consumers to participate in the market, thereby improving the overall efficiency and cost-effectiveness of grid operations [10]. These platforms not only promote more sustainable energy consumption but also enable a more equitable distribution of energy resources, as consumers can directly contribute to the stability of the grid.

Despite the promising advancements in AI-driven grid management, several challenges remain. One of the primary obstacles is the limited deployment of federated learning and decentralized optimization strategies at scale. Federated learning, which allows for data processing at the edge without sharing raw data, offers significant privacy benefits and efficiency improvements. However, it has yet to be widely implemented in large-scale grid systems due to technical and regulatory challenges. Similarly, decentralized optimization strategies, which aim to distribute decision-making across different nodes in the grid, are still in the early stages of deployment. These strategies require robust communication and coordination mechanisms, which have not yet been fully integrated into the existing grid infrastructure [12,11]. Overcoming these challenges will be crucial for realizing the full potential of AI in smart grids, ensuring that these systems are scalable, secure, and able to meet the demands of future energy networks.

Emerging AI Techniques and Their Role in Smart Grids

Next-generation AI technologies are transforming smart grids, pushing beyond traditional applications and offering innovative solutions to optimize energy distribution, improve grid reliability, and enhance system security. These advancements are enabling smarter, more adaptive, and more resilient energy networks. One of the most promising approaches in this regard is the integration of hybrid AI models, which combine neural networks with evolutionary algorithms. This combination has significantly improved the predictive accuracy of fault detection and renewable energy forecasting, addressing the inherent unpredictability of energy production from renewable sources like wind and solar. By leveraging the strengths of both neural networks for pattern recognition and evolutionary algorithms for optimization, hybrid models enhance the grid’s ability to forecast energy production and detect faults early, ensuring a more stable and reliable grid [12,18].

Another cutting-edge development is advanced probabilistic AI, particularly through the use of Bayesian Neural Networks (BNN). BNNs are designed to quantify uncertainty, providing a more robust approach to decision-making processes in renewable energy forecasting. These models allow grid operators to account for the inherent variability in renewable energy generation, such as fluctuations in wind speed or solar irradiance, by providing probabilistic outputs that capture uncertainty. This capability enhances decision-making, particularly in scenarios where precise predictions are difficult to achieve, and ensures that energy dispatch and grid balancing are optimized under uncertain conditions [14].

Additionally, Graph Neural Networks (GNN) are becoming increasingly important in AI-driven grid management. GNNs are particularly well-suited for capturing the complex relationships and structures within grid topologies, which are often represented as networks of interconnected nodes and edges. These graph-based models enable more accurate state estimation, helping grid operators determine the current state of the system with greater precision. GNNs are also effective at fault localization, allowing for faster identification of problem areas in the grid and facilitating quicker restoration times. Moreover, GNNs play a crucial role in cyberattack mitigation by detecting unusual patterns in grid behavior that may indicate malicious activities, such as cyberattacks targeting critical grid infrastructure. This ability to model and analyze grid topologies efficiently makes GNNs a valuable tool in ensuring the security and stability of modern smart grids [9].

Lastly, Deep Reinforcement Learning (DRL) is being applied to create self-learning grids. DRL systems are designed to optimize multi-objective tasks such as load balancing, market price determination, and power scheduling. By interacting with the grid environment and learning from the outcomes of its actions, DRL algorithms can dynamically adjust grid operations to meet multiple objectives, ensuring both operational efficiency and economic competitiveness. For example, DRL systems can learn to optimize energy distribution based on fluctuating demand and supply conditions, making real-time adjustments that improve grid flexibility and reduce energy waste. This self-learning capability allows the grid to adapt to changing conditions without human intervention, enhancing its resilience and operational efficiency [13].

These advancements in AI are setting the stage for the next generation of smart grids, which will be more adaptive, efficient, and secure. As these technologies continue to evolve, they will enable grids to better integrate renewable energy, respond to dynamic market conditions, and ensure a more sustainable and reliable energy future.

Challenges in AI-Powered Smart Grids

Despite the transformative potential of AI in enhancing the efficiency, reliability, and sustainability of smart grids, several significant challenges must be addressed to ensure its successful deployment and widespread adoption. These challenges can be broadly categorized into technical, economic, regulatory, ethical, and security concerns, each of which presents its own set of complexities.

Technical Challenges

One of the most prominent technical challenges in AI-powered smart grids is interoperability. AI models are often developed in isolation, and there is a lack of standardized protocols that would facilitate seamless integration into existing grid infrastructure. The diverse range of devices, systems, and technologies used in traditional grids and the newer smart grid components requires uniformity in communication standards and data formats. Without these standards, integrating AI-driven solutions with legacy systems becomes a time-consuming and complex task [5,6]. Furthermore, scalability remains a major issue. While AI models have demonstrated success in controlled environments and prototypes, transitioning these solutions from lab settings to large-scale, live grid deployments presents substantial technical hurdles. Smart grids involve complex, real-time data from multiple sources, and AI models must be able to handle the sheer volume and velocity of this data. Achieving scalability will require significant advancements in both AI algorithms and computational infrastructure to ensure that AI solutions can operate efficiently at the grid scale [12].

Economic and Regulatory Barriers

There are also several economic and regulatory barriers that hinder the adoption of AI in smart grids. Market restrictions present a key obstacle, especially when it comes to AI-enabled energy trading platforms. These platforms rely on real-time data and dynamic decision-making to optimize energy distribution and pricing, but their integration into traditional energy markets requires careful regulatory alignment. In many cases, energy markets are still governed by outdated regulations that do not account for the complexities and flexibility that AI technologies can bring to trading and distribution. Overcoming these regulatory hurdles is crucial for creating a more efficient and dynamic energy market [10]. Additionally, the lack of investment in AI-grid innovations remains a major barrier. While the potential benefits of AI are clear, the high upfront costs associated with deploying these advanced technologies often deter investment, particularly for distributed AI applications. The costs of installing the necessary infrastructure, including sensors, smart meters, and communication networks, combined with the expenses of developing AI models, can be prohibitive for many utilities and governments, especially in emerging markets [4,2].

*Ethical and Security Concerns
*
Lastly, ethical and security concerns must also be addressed to ensure the safe and equitable use of AI in smart grids. Data privacy is one of the most pressing ethical concerns, particularly with distributed AI systems like federated learning, which enable decentralized data processing across multiple edge devices. While federated learning preserves data privacy by ensuring that raw data is not shared between devices, it still involves the aggregation of sensitive data, which could be exploited if not properly secured. As AI systems in smart grids collect vast amounts of personal and operational data, safeguarding this data against unauthorized access and misuse is critical to maintaining consumer trust and complying with privacy regulations [12]. In addition to data privacy, cybersecurity threats pose a significant risk to the resilience of AI-powered control systems. The integration of AI in grid management makes the system more susceptible to adversarial attacks, where malicious actors may manipulate AI algorithms to disrupt grid operations. Ensuring that AI models are robust against such attacks is essential for maintaining the security and reliability of the grid. Protective measures, such as continuous monitoring, secure communication protocols, and adversarial training of AI models, are necessary to safeguard the grid from potential cyberattacks that could compromise its operation and stability [9].

In conclusion, while AI offers immense potential to revolutionize smart grid operations, addressing these technical, economic, regulatory, ethical, and security challenges is crucial to realizing its full potential. Only by overcoming these barriers can AI-powered smart grids become a mainstream solution for creating more efficient, resilient, and sustainable energy systems.

Future Directions

The future of AI applications in smart grids holds tremendous promise, with a number of exciting directions likely to shape the next phase of energy management. These advancements will further enhance the efficiency, reliability, and sustainability of grid systems, while also addressing some of the challenges currently faced by grid operators. Some of the key future directions for AI in smart grids include:

Autonomous Grid Management Systems

One of the most transformative developments in AI for smart grids is the move towards autonomous grid management systems. As AI technologies continue to evolve, the possibility of AI-driven microgrids that self-optimize operations without direct human intervention becomes increasingly feasible. These microgrids, which are localized networks of energy generation, storage, and distribution, will be capable of autonomously adjusting to real-time conditions such as changes in energy demand, renewable energy generation, and grid disruptions. AI systems will enable these microgrids to learn from past experiences and make real-time decisions regarding load balancing, fault detection, and energy dispatch. The goal is to create systems that not only manage their operations autonomously but also interact intelligently with the larger grid, contributing to overall system stability and reducing the need for manual oversight. This would significantly improve grid resilience, reduce operational costs, and enable faster response times during grid disturbances or peak demand periods.

Real-World Deployment of AI in Smart Grid Trials

Another significant future development is the real-world deployment of AI through large-scale pilot experiments and trials. As AI technologies mature, it will be essential to test their effectiveness in live grid environments. These large-scale pilot experiments will focus on testing multi-agent control systems and AI-powered market frameworks. Multi-agent systems (MAS), which involve multiple autonomous agents working together to manage distributed energy resources, will be tested in real-world trials to assess their ability to coordinate energy generation, storage, and consumption across various participants in the grid. Additionally, AI-powered market frameworks will allow for dynamic pricing, demand-side management, and energy trading based on real-time data and predictive analytics. These trials will provide critical insights into the scalability, efficiency, and economic viability of AI applications in real-world grid operations. Moreover, they will help identify potential regulatory and operational challenges, paving the way for more widespread adoption of AI in the energy sector [10].

Explainable AI (XAI) for Grid Decision-Making

As AI systems become more integrated into grid operations, the need for Explainable AI (XAI) will become increasingly important. Explainable AI refers to AI models and algorithms that provide transparent, interpretable explanations for their decisions, enabling grid operators and stakeholders to understand how and why certain decisions are made. This is particularly crucial in the context of grid decision-making, where automated AI-driven decisions impact everything from energy distribution to pricing and load balancing. Transparent AI models will foster greater trust in automated systems, particularly among stakeholders, regulators, and consumers who may have concerns about the “black-box” nature of many AI algorithms. By providing clear explanations for decisions, XAI can support regulatory acceptance, as it ensures that AI systems are operating in a manner that is both understandable and compliant with existing laws and policies. This transparency will be key to gaining widespread acceptance of AI-driven energy management solutions and ensuring that these technologies are deployed responsibly and ethically [6].

In conclusion, the future of AI in smart grids holds great potential for transforming energy management through the development of autonomous grid systems, large-scale pilot trials, and transparent AI models. These advancements will not only improve the efficiency and resilience of the grid but also contribute to a more sustainable and user-friendly energy landscape. However, continued research, development, and collaboration between industry stakeholders, governments, and consumers will be essential to realizing the full potential of AI in the energy sector.

Conclusion

In conclusion, Artificial Intelligence (AI) is fundamentally transforming the way smart grids operate by enhancing critical areas such as predictive maintenance, demand response, renewable energy integration, real-time data processing, and grid flexibility. AI applications are enabling smarter, more efficient, and adaptive grid systems that can autonomously adjust to changing conditions, optimize energy distribution, and improve the overall reliability and sustainability of the grid. Hybrid AI models and advanced reinforcement learning techniques are particularly driving innovation, offering significant improvements in energy forecasting, fault detection, and real-time decision-making. These AI technologies are enabling grids to predict and respond to fluctuations in energy demand, as well as optimize the use of renewable energy sources like solar and wind, which are inherently intermittent.

Despite the remarkable advancements, several challenges continue to hinder the wide-scale implementation of AI-driven solutions in smart grids. Scalability is one of the foremost challenges, as many AI models that have proven effective in controlled environments struggle to adapt when deployed in large-scale, real-time grid systems. These systems often require the processing of vast amounts of data, which presents significant computational and technical hurdles. Additionally, regulatory barriers pose challenges for the integration of AI technologies into existing market frameworks. AI-powered energy trading and demand response platforms, for example, require careful alignment with regulatory policies to ensure they comply with industry standards and legal frameworks. The data privacy concerns associated with distributed AI systems, such as federated learning, also remain a critical issue, particularly as smart grids handle sensitive consumer data. Ensuring the security and privacy of this data is essential for maintaining consumer trust and ensuring compliance with privacy laws and regulations [12,6].

Looking forward, continued research is crucial to address these challenges and advance AI technologies for smart grids. The focus should be on developing deployable AI solutions that can be scaled up for use in real-world grid environments while maintaining their effectiveness. Efforts should also be directed toward ensuring secure grid integration, which involves establishing standardized protocols for AI deployment and safeguarding systems from cybersecurity threats. Moreover, as AI models become more integral to grid decision-making, the development of explainable AI (XAI) will be essential for enhancing transparency and trust. Transparent and interpretable AI models will enable grid operators, regulators, and consumers to better understand AI-driven decisions, thereby fostering greater acceptance and facilitating the responsible deployment of these technologies [5,6,11].

In summary, while AI is poised to drive the next generation of intelligent, self-optimizing smart grids, overcoming the challenges related to scalability, regulatory alignment, data privacy, and model explainability is essential for unlocking its full potential. Through continued innovation and collaboration, AI will help shape a future of energy systems that are not only more efficient and sustainable but also more resilient and adaptable to the evolving needs of the energy sector.

References

M. Elkholy, O. Shalash, M. S. Hamad, and M. S. Saraya, “Empowering the grid: A comprehensive review of artificial intelligence techniques in smart grids,” 2024 International Telecommunications Conference (ITC-Egypt), 2024, pp. 513–518, doi: 10.1109/ITC-Egypt61547.2024.10620543.
A. Jaya, M. Rani, B. Kalpana, A. Srinivasan, S. Subramaniam, Msc Cybersecurity, S. A. Shiney, and V. Pandi, “Artificial intelligence-enabled smart grids: Enhancing efficiency and sustainability,” 2023 7th International Conference on Electronics, Communication and Aerospace Technology (ICECA), 2023, pp. 175–180, doi: 10.1109/ICECA58529.2023.10395590.
S. Kayan Kiliç, K. Özdemir, U. Yavanoglu, and S. Özdemir, “Enhancing smart grid efficiency through AI technologies,” 2024 IEEE International Conference on Big Data (BigData), 2024, pp. 7068–7074, doi: 10.1109/BigData62323.2024.10825117.
H. O. Buitrón-Barros, “Integración de inteligencia artificial en redes eléctricas inteligentes y su potencial transformador,” Horizon Nexus Journal, 2024, doi: 10.70881/hnj/v2/n2/37.
P. Arévalo and F. Jurado, “Impact of artificial intelligence on the planning and operation of distributed energy systems in smart grids,” Energies, 2024, doi: 10.3390/en17174501.
S. Ali and B. Choi, “State-of-the-art artificial intelligence techniques for distributed smart grids: A review,” Electronics, vol. 9, 2020, pp. 1030, doi: 10.3390/electronics9061030.
S. Lee, J. Seon, B. Hwang, S. Kim, Y. Sun, and J. Kim, “Recent trends and issues of energy management systems using machine learning,” Energies, 2024, doi: 10.3390/en17030624.
N. Noviati, S. D. Maulina, and S. Smith, “Smart grids: Integrating AI for efficient renewable energy utilization,” International Transactions on Artificial Intelligence (ITALIC), 2024, doi: 10.33050/italic.v3i1.644.
G. Arroyo-Figueroa, “An overview of applied artificial intelligence in power grids,” Int. J. Comb. Optim. Probl. Informatics, vol. 15, pp. 1–6, 2024, doi: 10.61467/2007.1558.2024.v15i4.532.
B. Eck, F. Fusco, R. Gormally, M. Purcell, and S. Tirupathi, “AI modelling and time-series forecasting systems for trading energy flexibility in distribution grids,” Proceedings of the Tenth ACM International Conference on Future Energy Systems, 2019, doi: 10.1145/3307772.3330158.
X. Wen, Q. Shen, W. Zheng, and H. Zhang, “AI-driven solar energy generation and smart grid integration: A holistic approach to enhancing renewable energy efficiency,” International Journal of Innovative Research in Engineering and Management, 2024, doi: 10.55524/ijirem.2024.11.4.8.
N. F. Akhter, A. Mia, and M. J. Talukder, “Python-based hybrid AI models for real-time grid stability analysis in solar energy networks,” Innovatech Engineering Journal, 2024, doi: 10.70937/faet.v1i01.24.
S. Pradeep, C. R. E. S. Rex, D. Kalaiyarasi, R. Dhandapani, M. Sakthivel, and K. Vijaipriya, “AI-driven fault detection in smart grids,” 2024 3rd Odisha International Conference on Electrical Power Engineering, Communication and Computing Technology (ODICON), 2024, pp. 1–6, doi: 10.1109/ODICON62106.2024.10797555.
A. Zafar, Y. Che, M. Sehnan, U. Afzal, A. D. Algarni, and H. Elmannai, “Optimizing solar power generation forecasting in smart grids: A hybrid convolutional neural network — Autoencoder Long Short-Term Memory approach,” Physica Scripta, 2024, doi: 10.1088/1402–4896/ad6cad.
T. Logenthiran, D. Srinivasan, and T. Shun, “Demand side management in smart grid using heuristic optimization,” IEEE Transactions on Smart Grid, vol. 3, pp. 1244–1252, 2012, doi: 10.1109/TSG.2012.2195686.