Next Article in Journal
Novel Real-Time Power System Scheduling Based on Behavioral Cloning of a Grid Expert Strategy with Integrated Graph Neural Networks
Previous Article in Journal
Determination and Verification of Real-Time Transient Stability of Jeju System According to Increase in Renewable Energy
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Artificial Intelligence in Renewable Energy Systems: Applications and Security Challenges

1
State Grid Information and Telecommunication Group Co., Ltd., Beijing 100029, China
2
School of Automation Science and Engineering, Xi’an Jiaotong University, Xi’an 710049, China
3
National Innovation Platform (Center) for Industry-Education Integration of Energy Storage Technology, Xi’an Jiaotong University, Xi’an 710049, China
*
Author to whom correspondence should be addressed.
Energies 2025, 18(8), 1931; https://doi.org/10.3390/en18081931
Submission received: 3 March 2025 / Revised: 5 April 2025 / Accepted: 8 April 2025 / Published: 10 April 2025
(This article belongs to the Section A: Sustainable Energy)

Abstract

:
In the context of the global active pursuit of sustainable development and the heightened priority placed on sustainable energy, renewable energy systems, as a crucial solution to energy crises and environmental challenges, are of increasing significance. The extensive development and utilization of renewable energy sources such as wind and solar have become the core driving force for promoting the transformation of the energy structure. The research and construction of energy storage systems have also become trends in future energy development. AI, with its powerful data-processing and intelligent decision-making capabilities, has been deeply integrated into multiple key aspects of renewable energy systems. This review fills a gap in the relevant literature by conducting an updated technological assessment of the application of AI technology in renewable energy systems including wind power systems, PV power systems, energy storage systems, and others. Moreover, this paper analyzes the security challenges of AI in renewable energy systems. The primary aim of this review is to identify the advantages and existing security challenges of introducing AI technology into renewable energy systems, so as to help improve the production efficiency and information security level of different forms of renewable energy systems.

1. Introduction

1.1. Research Background and Motivation

In the contemporary global context, the imperative for sustainable development and the vigorous pursuit of sustainable energy have gained unprecedented momentum. This shift is driven by the urgent need to address energy crises and environmental challenges, which have become increasingly pressing in the face of climate change and resource depletion. Renewable energy systems, encompassing wind, solar, hydropower, nuclear, hydrogen, and geothermal energy, are emerging as pivotal solutions to meet the growing demand for clean and efficient energy sources. These systems not only offer a sustainable alternative to traditional fossil fuels but also contribute significantly to reducing greenhouse gas emissions and mitigating the impacts of climate change.
The integration of renewable energy sources into the global energy mix presents both opportunities and challenges. On one hand, the intermittent nature of wind and solar energy, coupled with the geographical dispersion of renewable resources, necessitates innovative solutions to ensure a stable and reliable power supply. On the other hand, the rapid advancement of AI and ML technologies offers a promising avenue to enhance the efficiency, reliability, and integration of renewable energy systems. AI has the potential to revolutionize various aspects of renewable energy, from power generation and prediction to system optimization and fault diagnosis.
This review paper aims to provide a comprehensive overview of the applications and security challenges of AI in renewable energy systems. We will explore how AI technologies are being utilized to optimize the performance of wind power, PV systems, energy storage systems, and other renewable energy technologies. Additionally, we will discuss the security implications of integrating AI into these systems, highlighting the potential risks and vulnerabilities, as well as the strategies and solutions to address them. The paper will also review the current state of research in this field, identify gaps and future research directions, and provide insights into the practical implementation of AI-driven solutions for a more sustainable and secure energy future.

1.2. Materials and Methods

To ensure the comprehensiveness and relevance of this review, a systematic literature search was conducted. The search was carried out in multiple academic databases, including Web of Science, IEEE Xplore, and ScienceDirect. The keywords used in the search included “artificial intelligence”, “renewable energy systems”, “wind power systems”, “photovoltaic power systems”, “energy storage systems”, “security challenges”, and their combinations. These keywords were carefully selected to cover the broad scope of the research topic. For the screening criteria, only peer-reviewed journal articles and conference papers published in the past 15 years (2010–2025) were considered. Articles that were not directly related to the application of AI in renewable energy systems or did not discuss security challenges were excluded. This process ensured that the literature included in this review was up-to-date, relevant, and of high quality.
The organization of the paper is as follows: Section 1 is a brief introduction of the whole review. Section 2 elucidates the mechanisms through which AI facilitates the operation of wind power systems. It further explores the regional applications of AI within the wind power domain and analyzes the advantages conferred by AI implementation, as well as the security challenges of AI in wind power systems. Section 3 presents the applications of AI in PV power systems. It also examines how AI has transformed the traditional operational paradigms of PV power systems, and the security challenges that AI brings. Section 4 elaborates on the integration of energy storage systems and AI. Additionally, it analyzes the resultant advantages and security problems brought by this synergy. Section 5 introduces the applications and security risks of AI in some other energy systems. Section 6 draws concise conclusions based on the research conducted throughout the paper.

2. Wind Power Systems

2.1. Background

Against the backdrop of the global community’s active response to climate change and the vigorous promotion of energy transformation, the demand for clean energy is growing rapidly. Wind power, with its numerous advantages such as renewability, zero pollution, and abundant resources, has become a mainstay in the field of renewable energy, and its development has attracted significant attention from all sectors [1,2]. In recent years, with the continuous progress of intelligent technologies, AI has gradually permeated all aspects of wind power systems, from wind speed prediction [3] to fault diagnosis [4], and optimized the scheduling [5] of wind turbines.
The application of AI not only enhances the efficiency of wind power generation but also promotes the intelligent development of renewable energy. However, the widespread application of AI has also brought some security risks, especially in data privacy [6], system stability, and hardware security. To visually present the application scenarios, algorithm models, security issues, and corresponding solutions of AI in the field of wind power generation, Figure 1 summarizes its technical architecture and core components from multiple perspectives. This chapter focuses on exploring the application and security issues of AI in wind power systems and analyzes corresponding solutions.

2.2. Application of AI in Wind Power Systems

The stability and efficiency of wind power generation rely heavily on accurate wind power prediction. Traditional wind power prediction methods include physical methods based on NWP data and time series statistical methods based on historical data [7,8]. Physical methods calculate the output power by solving the physical equations of wind speed and combining them with the power curve of wind farms. However, they suffer from issues such as high computational complexity, low spatial resolution, and errors in meteorological data, and are mostly used for medium- and long-term predictions. Simple time series methods like the persistence method have certain accuracy in ultra-short-term wind power prediction when the data change smoothly, but they have significant overall limitations.
Neural networks in ML algorithms, with their powerful nonlinear mapping capabilities, have demonstrated excellent performance in the field of wind power prediction [9]. These networks can deeply explore the complex relationships among historical meteorological data, geographical information, and wind farm operating data to build high-precision prediction models. The LSTM, a specialized type of recurrent neural network, exhibits outstanding abilities in processing time series data. It effectively captures long-term dependencies and complex patterns in wind speed time series, thereby enhancing prediction accuracy. Reference [10] introduces a short-term wind power prediction method for wind farms based on a hybrid approach combining LSTM and NARX neural networks. In this model, LSTM is applied to make short-term predictions of wind speed time series while avoiding gradient problems. The LSTM output then serves as input for the NARX network, establishing a hybrid model that predicts wind power for the next 48 h, which effectively enhances short-term prediction precision. Reference [11] addresses the convergence and local extremum challenges found in traditional Sparrow Search Algorithm-optimized CNN-BiLSTM power prediction models. The authors implement refraction inverse learning, sine-cosine, and Cauchy mutation strategies to improve the algorithm and construct an ISSA-CNN-BiLSTM model. Experimental validation using data from Inner Mongolia wind farms demonstrates that this model achieves higher prediction accuracy compared to the SSA-CNN-BiLSTM and CNN-BiLSTM models, confirming its efficacy in wind power forecasting. Additionally, Reference [12] proposes a wind power prediction approach based solely on the LSTM neural network. After processing and validating the model with data from a Spanish wind farm, the results reveal that this method delivers more accurate predictions than traditional algorithms.
Compared with traditional wind power prediction methods, AI-based methods show progress in multiple aspects. Traditional physical methods rely on NWP data, suffering from high computational complexity, low spatial resolution, and susceptibility to meteorological data errors, and are mainly applicable to medium-and long-term predictions. Simple time-series statistical methods like the persistence method have certain accuracy in ultra-short-term predictions when data change smoothly but have significant overall limitations. In contrast, AI methods such as LSTM and NARX in neural networks can deeply explore the complex relationships among historical meteorological data, geographical information, and wind power operation data with their strong nonlinear mapping capabilities. They can build high-precision prediction models, significantly improving prediction accuracy and performing well in predictions at different time scales. In terms of comprehensive applications, AI-based combined prediction methods, whether they are weight-coefficient-based or fusion-based, enhance the robustness and accuracy of predictions by integrating the advantages of different methods. Moreover, AI methods can effectively handle complex nonlinear problems and automatically extract data features, which is beyond the reach of traditional methods. Overall, AI methods bring more accurate, intelligent, and adaptable solutions to wind power prediction, strongly promoting the efficient and stable operation of wind power systems.
Additionally, combined prediction methods have shown continuous evolution in this field. Weight coefficient-based combined prediction determines reasonable weight coefficients by integrating the independent prediction results from different methods, such as the equal weight method and least squares method, which enhances the robustness of prediction outcomes. In contrast, fusion-based combined prediction employs different methods at various stages of the prediction process. For instance, techniques such as wavelet transform and empirical mode decomposition are utilized to decompose time series before prediction, followed by result combination. Furthermore, genetic algorithms are employed to optimize model parameters, where prediction residuals undergo estimation and correction. These approaches have collectively improved wind power prediction accuracy to a significant extent [13,14,15].
AI technology has brought a revolutionary change to the control system of wind turbines, enabling intelligent and autonomous regulation of wind turbines. Reinforcement learning algorithms allow wind turbines to automatically and dynamically optimize blade angles and rotational speeds based on multi-dimensional environmental variables such as real-time wind speed, wind direction, and temperature. This ensures that wind turbines can maintain optimal power generation efficiency under various operating conditions [16,17,18]. This adaptive control strategy enables wind turbines to flexibly respond to complex and variable external environmental conditions, achieving more accurate wind turbine control. In addition, intelligent control effectively reduces mechanical wear and fatigue damage during the operation of wind turbines, thereby extending the service life of wind turbines and significantly reducing the full life cycle cost of equipment [19].
In the field of wind power generation, fault diagnosis of wind turbines is crucial for ensuring the reliable operation of wind turbines and reducing operation and maintenance costs. AI algorithms play a key role in this regard [20]. Traditional signal processing methods such as vibration analysis were once the mainstream. For example, Reference [21] uses vibration acceleration signals and generator power for statistical processing to analyze the behavior of wind turbine bearings under non-stationary conditions. The data are decomposed into different load sub-intervals to assist in fault diagnosis. Reference [22] adopts the EMD combined with the Ensemble Empirical Mode Decomposition method to demodulate and analyze the vibration signals of planetary gearboxes, and compares the Fourier spectra of the demodulated signal envelopes with theoretical values to detect gearbox faults. However, these methods have limitations, such as limited processing capabilities for complex nonlinear data and difficulty in automatically extracting features from large amounts of data.
ML algorithms have brought new breakthroughs to the field of fault diagnosis. In terms of classification techniques, Reference [23] applies the SVM to analyze the SCADA data and power curves of wind turbines, and to screen and diagnose faults. It performs well in the early fault prediction of multiple components but does not solve the problems of multi-classification and feature selection. Reference [24] uses the RF combined with PCA to identify faults of multiple components of wind turbines and determine key SCADA signals. The training is efficient and can handle large datasets. References [25,26] uses regression trees to predict the power output of wind turbines based on parameters such as wind speed, which is more accurate than traditional curve fitting and can be applied to new data. Reference [27] processes SCADA data for anomaly detection based on the Pearson correlation coefficient and self-organizing maps. Although effective, it has limitations due to its dependence on power curves.
With the development of AI technology, DL has been widely used in the field of wind turbine fault diagnosis in recent years. Reference [28] uses an ANN to predict the percentage of the remaining life of wind turbine components based on SCADA data, but does not explain the basis for network training and architecture selection. The ELM model proposed in Reference [29] is faster than traditional feedforward neural networks in fault recognition, but the model details and comparative analysis are lacking. RNN and their variants such as LSTM can process time series data. Reference [30] uses LSTM for wind speed prediction for wind turbine fatigue analysis, and Zhu et al. [31] and Liu et al. [32] use LSTM for wind power prediction, showing a higher accuracy than traditional methods. Reference [33] focuses on the health prediction of wind power equipment based on RNN. By using multiple memory cell variants like LSTM, BiLSTM, and GRU, and adjusting the training hyperparameters via a Bayesian optimization loop, the prediction accuracy has been significantly enhanced, offering crucial reference for the development of wind power fault diagnosis tech.
Existing research shows that artificial intelligence-based fault diagnosis methods are significantly improved compared to traditional diagnosis methods, in terms of indicators such as accuracy, recall rate, and F1-score. A comparative experiment using the Case Western Reserve University Bearing Dataset shows that, under the same vibration signal samples, the detection accuracy of an ANN-based algorithm for inner-race faults of bearings reaches 97.2%, which is significantly higher than the 84.5% of the traditional envelope analysis method. Research based on the SCADA data of the National Renewable Energy Laboratory further verifies that the F1-score of the LSTM model in detecting abnormal gearbox temperatures is 23% higher than that of the Pearson correlation coefficient method, and the false alarm rate is reduced to 1.8%. In terms of scalability, a deep model running on the TensorFlow framework can achieve minute-level parallel diagnosis for gigawatt-scale wind farms on the AWS cloud platform. In contrast, traditional vibration analysis systems are limited by local computing resources and take several hours to process the same amount of data. Overall, through automatic feature extraction, end-to-end learning, and a distributed computing architecture, artificial intelligence technology has comprehensively surpassed traditional methods in three dimensions: diagnostic accuracy, complex pattern recognition, and adaptability to large-scale systems, marking the entry of wind power operation and maintenance into an intelligent decision-making era driven by data.

2.3. Security Challenges and Solutions of AI in Wind Power System

As the position of wind power in the global energy structure gradually rises, how to ensure the security and reliability of wind power systems has become an important issue that needs to be addressed urgently. The security issues of wind power systems involve multiple aspects such as grid stability, equipment fault monitoring, and data privacy protection. Research shows that with the increase in wind power penetration, voltage security issues in the power grid become particularly prominent. To this end, Gao et al. [34] propose a dual-time-scale reactive power optimization control strategy. By coordinating voltage regulation equipment and reactive power compensation equipment, it effectively reduces voltage fluctuations and ensures the voltage security of wind power systems. In addition, Margaris et al. [35] study the dynamic security issues of isolated power systems under high wind power penetration, analyze the impact of the fault ride-through capability of wind turbines on system frequency and load shedding, and provide a theoretical basis for grid security.
On the other hand, the abnormal detection of wind power equipment is also an important part of ensuring system security. Due to the distributed characteristics of wind power equipment far from cities, traditional fault diagnosis methods cannot meet the needs of real-time monitoring. Ding et al. [36] propose an abnormal detection system for wind power equipment based on CNN, which can effectively identify equipment faults through real-time data analysis and improve the intelligent level of equipment operation and maintenance. Xu et al. [37] design an intelligent dynamic security assessment framework, combined with ELM technology, to conduct real-time dynamic security assessments of large-scale wind power systems, providing a guarantee for the safe operation of wind power systems. Zhang et al. [38] study the IIV phenomenon in the dynamic security assessment of wind power systems, reveal the potential risks of IIV to system security, and call for improving existing dynamic security assessment methods to address this challenge.
With the increasingly severe issues of data privacy and security, how to conduct efficient wind power data prediction and analysis while ensuring privacy has become an important research direction. Ahmadi [6] proposes a wind power prediction method based on federated learning, which improves the prediction accuracy and generalization ability while protecting data privacy. In addition, the agile authentication mechanism proposed by Shu et al. [39] effectively ensures the communication security of wind power monitoring systems. Combined with the power grid fault risk assessment method based on dynamic probabilistic power flow proposed by Zhao et al. [40], which uses fuzzy reasoning for security early warning, these studies provide important technical support for wind power systems in ensuring data privacy, optimizing predictions, and enhancing security, promoting the sustainable development of intelligent wind power systems.

3. PV Power Systems

3.1. Background

PV power systems have the advantages of being clean and renewable. However, they also face numerous challenges. For example, the power generation capacity is greatly affected by external factors such as weather, resulting in unstable output. Meanwhile, the operation, maintenance, and fault management of PV systems are relatively complicated.
In recent years, scholars have applied AI technology to PV power systems with significant outcomes. Advanced models such as CNN and LSTM have demonstrated remarkable accuracy in predicting PV system output. Concurrently, image recognition technology deployed via unmanned aerial vehicles has enabled efficient fault detection and rapid problem localization in PV modules. These intelligent approaches not only enhance operation and maintenance protocols for PV systems but also establish a reliable foundation for optimizing power grid dispatching and improving system stability.
Furthermore, the implementation of AI technology in the security protection of PV power systems—encompassing equipment condition monitoring, personnel behavior analysis, and network security defense—lays the foundation for developing more secure and efficient intelligent PV systems.
This chapter provides a comprehensive review of AI applications in PV power plants. It critically examines innovative implementations in real-time fault detection and diagnosis, adaptive power forecasting under weather fluctuations, intelligent operation and maintenance systems, and grid-responsive energy management, while summarizing current technological advancements. AI-driven approaches demonstrate marked improvements in operational efficiency and accuracy compared to conventional methods, particularly in handling complex data patterns and dynamic environmental conditions. The integration of AI extends to cybersecurity enhancement, addressing vulnerabilities from data noise and potential cyberattacks through adaptive learning and explainable decision-making frameworks. Through systematic analysis of the existing literature, this chapter offers theoretical references and practical guidance for the intelligent development and security enhancement of PV power systems. Figure 2 illustrates the applications of AI in PV systems, associated security challenges, and corresponding algorithmic technologies.

3.2. Application of AI in PV Power Systems

AI techniques such as CNN, RF, and SVM have demonstrated superior performance over traditional fault detection methods in terms of speed and accuracy. Taghezouit et al. [41] used DL technology for fault diagnosis of PV systems. They developed a multi-channel time series data model using CNN to monitor new energy power generation systems, including PV systems, in real time. This model processes a large amount of multi-dimensional data, such as current, voltage, power, etc. By learning normal and abnormal patterns, it can accurately detect faults in PV modules and inverters, highlighting the advantages of CNN in handling complex data and improving the accuracy of fault detection. Lukas et al. [42] proposed the data-driven Fault Detective algorithm, which utilizes RF regression to adaptively model system behavior and detect anomalies. This method outperforms fixed threshold-based approaches by accommodating diverse system layouts and sensor configurations.
AI excels in identifying specific faults such as hot spots and inverter failures. For instance, Jia et al. [43] applied ANN models to analyze thermal infrared images, achieving effective hotspot detection in PV modules. Christian et al. [44] focused on the application of explainable AI technology in PV fault detection and evaluated the performance of different detection instruments. The study used models such as decision trees, RF, and SVM, which can not only detect faults in PV systems but also provide detailed fault explanations, which is crucial for maintenance personnel to quickly locate and repair faults.
Practical applications emphasize efficiency gains. Zhounan et al. [45] explored PV project safety management based on drone AI inspections. This technology integrates drones with AI algorithms, where drones equipped with multiple devices capture images, and the data are transmitted to the cloud in real-time. AI then analyzes anomalies and generates report alerts. Polymeropoulos et al. [46] highlighted drone-based CNN inspections, which use high-resolution cameras mounted on drones to inspect PV power plants from the air. This method can quickly locate potential faults, improve detection efficiency, and reduce the cost and risk of manual inspections.
Solar irradiance, temperature, and cloud cover are critical weather parameters for PV output prediction. Mansouri et al. [47,48,49] explored the application of AI techniques in PV power data mining and forecasting, focusing on meteorological and spatiotemporal features. They employed CNN and LSTM models to analyze meteorological and historical power generation data, identifying key factors for short-term and ultra-short-term forecasting. This approach improved accuracy and efficiency, optimized power plant performance, reduced output uncertainty, and enhanced system stability while supporting grid integration. Hybrid AI architectures improve robustness to weather variability by integrating real-time data and network topology features. For instance, Shao et al. [50] developed an ultra-short-term PV power forecasting method based on DWT-CNN-LSTM, while Xiang et al. [51] proposed a short-term forecasting model using an efficient channel attention mechanism-optimized TCN-GRU neural network. Zhang et al. [52] explored a real-time PV consumption evaluation method for distribution networks based on GAT, leveraging its advantages in processing topology and node features. By combining topology mining features, they developed two models to evaluate node consumption capacity and establish constraints. Simulations confirmed that this method performs well in terms of accuracy, real-time processing, and generalization ability for both known and unknown topologies.
Seasonal and geographic variations are addressed through adaptive training. Chen et al. [53] proposed a distributed PV power data forecasting method based on Bayesian neural networks, constructing an improved model that adapts its structure according to different environmental characteristics. The model processes weather forecasts and historical power generation data, with dimensional reduction of weather data before input, generating a PV output distribution that improves accuracy and provides reliable references for grid scheduling. Amadou et al. [54] focused on the tropical rainy and dry seasons, studying the changes in the PV system power output and optimizing them using AI algorithms to enable coordinated operation with energy storage, charging during sunny days, and discharging during cloudy periods to ensure stable energy supply. Ge et al. [55] investigated PV power forecasting based on the PSO-BP algorithm, highlighting its advantages in handling complex data and extracting patterns. This algorithm employs particle swarm optimization to initialize the weights and thresholds of BP neural networks, addressing the local optimum and slow convergence issues of traditional BP networks. Feng et al. [56] reviewed combined XGBoost–LSTM models, demonstrating high accuracy under varying weather conditions. Additionally, researchers have explored other AI applications in PV systems.
AI enhances grid integration through accurate forecasting and energy storage optimization. Al et al. [57] explored the application of AI control in PV energy management, using ANFIS for the coordinated optimization of PV and energy storage systems under different climatic conditions. They also modeled and managed lithium-ion batteries and utilized feedforward neural networks, ANN, and ANFIS to predict performance, adjusting outputs in real time and scheduling in advance for efficient energy usage. In energy management, AI technologies are used to control and coordinate multiple PV inverters to achieve reactive power and voltage control. For instance, using the MAAC framework, PV inverters are treated as agents, coordinating their reactive power generation or absorption according to system requirements, maintaining voltage within ±5%, reducing power losses, and improving voltage controllability [58].
Case studies highlight operational advancements: Some studies have proposed predictive methods based on digital twin technology, creating virtual replicas of physical systems to reflect and predict their state and performance in real time [59]. Other studies have built BP neural network and wavelet neural network models using neural network algorithms to analyze historical and real-time data for accurate prediction of PV power output [60].

3.3. Security Challenges and Solutions of AI in PV Power System

Weather-induced data inconsistencies, such as noise and missing values, challenge AI reliability. Bakht et al. [49] emphasized the impact of meteorological uncertainty on prediction accuracy. Mandal et al. [61] applied wavelet transforms to denoise weather-affected PV data. Due to the high impact of weather on PV power, making it unpredictable and threatening grid stability, developing reliable forecasting algorithms to reduce errors is crucial for efficient integration of variable energy resources into the grid. Accurate PV power forecasting is essential. Salazar-Peña et al. [62] used synthetic datasets to enhance ANN generalization. They proposed an AI-based dynamic fault detection and performance analysis method for PV systems. This study used Python 3.11’s PVlib library and dynamic loss quantification algorithms to process meteorological, operational, and technical data. ANN were trained on synthetic datasets to simulate actual PV system failures. The method achieved an average accuracy of 82.2% and a maximum accuracy of 92.6% in fault detection models.
AI-driven security frameworks mitigate cyber threats. Kim et al. [63] proposed a multi-layered defense system for PV-rich grids, which is a novel method to enhance the network security and resilience of solar photovoltaic-rich distribution systems under potential cyber-attacks. This method focuses on the secure operation post-attack, providing multi-layered security frameworks combined with various technologies for comprehensive protection. Murat et al. presented methods using XAI tools to forecast solar PV power generation, which improves transparency in fault diagnosis and attack detection. Although AI has many benefits in smart grid applications, it often has a “black-box” nature, especially in PV forecasting, which involves multiple parameters. XAI, as an emerging field, helps to understand AI decision-making processes and predictions [64].

4. Energy Storage Systems

4.1. Background

With the global energy transition and the rapid development of renewable energy, optimizing energy efficiency, promoting environmental sustainability, and achieving seamless integration with existing power systems have become key areas of research [65,66]. In this process, the study of smart grids has gradually become a central focus. Particularly in the face of the inherent challenges of renewable energy, such as the geographical distance between renewable energy sources and consumption points, the limitations of weather forecasting, and the continuous adjustment of power demand, the role of energy storage systems has become increasingly crucial [67]. Energy storage technology plays a key role not only in capturing and accumulating energy but also in regulating the power generation output of renewable energy, enhancing the flexibility, rapid response capability, and frequency control functions of power systems, and thus improving the feasibility of large-scale renewable energy applications [68,69,70]. Moreover, energy storage systems can provide backup power during emergencies or power outages, further ensuring the stability of the power supply [71].
However, with the development of renewable energy, the complexity of energy storage systems and their dynamic interactions with the grid, power generation, and load demand have posed significant challenges for the optimization, scheduling, and management of energy storage systems. AI, as an advanced data-driven technology, is gradually becoming an effective tool to address these issues, helping to improve the operational efficiency, reliability, and economic viability of energy storage systems. AI can play a role in various aspects of energy storage systems, including demand load forecasting [72], battery health management [73], and the optimization and scheduling of energy storage systems [74]. This section will primarily introduce the application of AI in energy storage systems and analyze some of the security issues that exist [75]. The applications of AI in energy storage systems and the related security issues, along with the corresponding algorithmic technologies, are shown in Figure 3.

4.2. Application of AI in Energy Storage Systems

The application of AI in energy storage systems starts with demand load forecasting in smart grids. Initially, load forecasting relied on traditional statistical techniques, time series analyses, and simple predictive models [76]. However, as energy demand grows and its complexity increases, more advanced technologies are required to address the challenges of grid variability [77,78]. Traditional methods, such as ARIMA and Support Vector Regression, achieve a certain level of accuracy in short-term load forecasting. However, their predictive capability is limited when dealing with complex nonlinear energy consumption patterns. The emergence of DL has significantly transformed the approach to load forecasting in smart grids. DL can automatically extract features and adapt to changing situations, making it highly suitable for the dynamic aspects of energy usage. The development of complex models such as LSTMs and RNNs has led to significant progress in estimating and forecasting changes in energy usage over time [79,80]. DL technologies can solve complexity, handle nonlinear problems, and provide accurate probabilistic load forecasts. These DL technologies offer advantages such as feature selection, capturing complex patterns, and providing uncertainty estimates, all of which are highly valuable for energy management and decision-making in smart grids. For example, LSTM are adept at handling time series data and can capture long-term dependencies in historical data, making them suitable for the seasonal, cyclical, and trend variations in electricity load. By analyzing multidimensional information such as historical load, climate, and holidays, the LSTM model can make high-precision predictions for future load [81,82]. Widodo et al. used LSTM to forecast demand loads for renewable energy power plants [82], which accurately predicted electricity consumption and renewable energy generation, improving energy management services and renewable energy distribution. It can effectively handle complex large-scale data structures, making it suitable for forecasting smart grid systems with uncertainty and nonlinearity. These DL architectures have demonstrated outstanding capabilities in handling temporal dependencies, enabling more precise load pattern predictions, especially in dynamic and nonlinear connections [83]. This can help energy storage systems gain a comprehensive understanding of the system power demand, providing sufficient information for subsequent scheduling and other processes.
Secondly, AI has been widely researched and applied in battery health management. Battery health management is essential for ensuring the normal operation of energy storage systems and extending the lifespan of batteries. It typically requires monitoring the battery’s health SOC to ensure the battery’s well-being [84]. The health of a battery is often evaluated through indicators such as charging efficiency, internal resistance, and capacity [85]. Traditional battery SOH evaluation methods are usually based on a single indicator or empirical rules. However, due to the complex usage environment and load conditions of batteries, these methods are often unable to meet the precise evaluation requirements [84]. AI can analyze historical charge/discharge data, current electrical parameters, and environmental conditions, and use ML algorithms to establish battery SOH prediction models. These models can assess the battery’s health in real time, predict its remaining useful life, and provide a basis for battery maintenance and replacement [86]. The SOC indicates the percentage of the battery’s remaining charge. The accuracy of SOC estimation is crucial for the operation of energy storage systems, as it not only affects the battery’s charging and discharging strategies but also influences its overall lifespan. ML algorithms can integrate multidimensional data such as voltage, current, and temperature under various operating conditions to achieve more accurate SOC estimation. AI methods, through adaptive learning, can effectively compensate for errors caused by factors such as battery aging, ensuring that the battery operates in its optimal state and extending its service life [87].
Additionally, AI plays a crucial role in the optimization and scheduling of energy storage systems. One of the key highlights of AI applications is the optimization of charging and discharging strategies for energy storage systems [88,89]. AI can optimize these strategies based on load forecasting results and the characteristics of the energy storage system. By using reinforcement learning algorithms, such as DQN, the state of the energy storage system (including the battery’s SOC, current load, electricity prices, etc.) can be inputted, with the goal of optimizing long-term operational benefits (such as cost minimization and profit maximization). The algorithm learns the best charging and discharging strategies [90]. This not only improves the efficiency of energy storage resources but also maximizes the economic benefits of the energy storage system. Through intelligent scheduling, the system can provide support during times of insufficient supply or peak demand, optimizing the overall operation of the power system. Furthermore, in systems with multiple energy storage devices, AI can optimize power distribution [91]. By employing a multi-agent system, each agent represents an energy storage device. Through communication and collaboration among these agents, optimal power distribution can be achieved [92]. For example, in a system with different batteries, a multi-agent system can reasonably allocate power based on the current state of each battery, its performance characteristics, and the grid’s power demand. This optimizes the performance of the entire energy storage system, extends the service life of the storage devices, and meets the grid’s power regulation requirements [93]. Compared to traditional optimization scheduling techniques, reinforcement learning-based methods can adaptively adjust charging and discharging strategies in dynamic environments, improving the economic efficiency and operational stability of energy storage systems. Particularly in energy storage management problems with high uncertainty, AI demonstrates more significant advantages.
Recently, research on the development of efficient energy storage devices using artificial intelligence technology has been worth paying attention to. Carvalho et al. [94] developed a methodology combining artificial intelligence with quantum mechanics to accelerate the discovery of suitable organic materials for lithium-ion battery cathodes, which through high-throughput screening identified 459 promising organic electrode materials with the potential to achieve theoretical energy densities exceeding 1000 Wh kg−1. Citroni et al. [95] noted that ML can be used to predict energy source availability, manage harvested energy, optimize power consumption in sensor nodes, and improve material selection for energy harvesting devices. Reference [96] reviewed how artificial intelligence techniques can bridge nano- and microscale X-ray tomography for battery research, demonstrating how AI and machine learning can be leveraged for image segmentation, analysis, and the development of multiphysics and multiscale predictive models, providing new pathways for battery development and optimization. Qiu et al. [97] examined the application of multiscale computational methods and artificial intelligence models in analyzing the electrochemical performance of Li-ion battery materials, starting from the relationship between energy and variable parameters to systematically explain how theoretical calculations can predict energy density, rate capability, and cycling stability, providing methodological references for Li-ion battery material design and suggesting that the deep combination of multiscale computations, experimental data, and machine learning will become a more powerful tool for discovering new Li-ion battery materials. Li et al. [98] proposed an artificial intelligence-based design approach for power converter circuit parameter design, achieving automation in analysis, deduction, and optimization processes through batch-normalization neural networks and genetic algorithms, which not only reduces engineers’ workload but also improves design accuracy, with its feasibility and high precision validated in the design of a synchronous buck converter for a 48 V to 12 V accessory-load power supply system in electric vehicles.

4.3. Security Challenges and Solutions of AI in Energy Storage Generation

However, although AI has many applications in energy storage systems and has improved their operational efficiency, the widespread adoption of AI inevitably brings some security concerns [99,100].
The first issue is data security [101]. In energy storage systems, AI needs to collect large amounts of data, such as operational parameters of storage devices (voltage, current, temperature, etc.), user behavior patterns, and energy consumption data [102]. These data are stored in databases, and if the AI system’s cybersecurity measures are inadequate, hackers may infiltrate the database and steal this sensitive information [103,104]. If hackers gain access to the energy usage habits of storage system users, they could infer the user’s daily routines, which may pose a threat to the user’s privacy and security [105]. Additionally, the data from energy storage systems are crucial for energy suppliers and grid operators. A data breach could lead to the loss of trade secrets and negatively impact the company’s competitiveness. Moreover, malicious attackers may tamper with the data used by AI to monitor and control the energy storage system [106]. If data from temperature sensors in energy storage devices are altered, the AI system might make erroneous decisions based on incorrect data, such as failing to trigger the cooling system when the actual temperature is too high. This could lead to the overheating of storage devices and potentially cause safety incidents, such as fires.
There are also algorithm security issues [107]. AI algorithms may have vulnerabilities or inaccuracies [105]. In the energy management of storage systems, if AI-based predictive algorithms do not adequately account for the impact of weather changes on renewable energy generation (such as solar and wind energy), then they may schedule charging and discharging plans for storage devices incorrectly. If the predicted energy generation is higher than the actual generation, the storage devices may discharge excessively, resulting in insufficient power supply to the grid when energy support is needed. Furthermore, if the data used during the algorithm’s training process are incomplete or biased, it can lead to a decline in the algorithm’s performance [108]. In algorithms used for predicting battery lifespan in energy storage systems, if data from various complex operating conditions are not included, the predicted battery life may significantly differ from the actual lifespan. This could lead to the battery failing earlier than expected, increasing the risk of energy storage system failures.

5. Other Renewable Energy Systems

5.1. Background

Amid global efforts to mitigate climate change and accelerate the transition to sustainable energy, the demand for clean, efficient, and renewable energy sources is rapidly increasing. Technologies such as hydropower, nuclear energy, hydrogen energy, and geothermal energy have emerged as vital pillars of the global energy mix, each offering unique advantages and attracting widespread attention. In recent years, the advancement of AI has revolutionized these sectors by enabling innovations across diverse areas, including the optimization of energy production, predictive maintenance, anomaly detection, and system integration. From improving hydropower reservoir operations and enhancing nuclear reactor safety to optimizing hydrogen production processes and advancing geothermal exploration, AI has demonstrated its transformative potential in driving operational efficiency and sustainability. However, alongside these advancements, the widespread adoption of AI introduces significant challenges, including cybersecurity vulnerabilities, data privacy risks, and system stability concerns. This section explores the application of AI across these energy domains, highlights the associated risks, and analyzes strategies to address these challenges, ensuring the secure and sustainable development of next-generation energy systems. The applications of AI in other renewable energy systems, related security issues, and corresponding algorithmic techniques are shown in Figure 4.

5.2. Hydropower Systems

AI has revolutionized the hydropower sector by optimizing energy production, improving operational efficiency, and integrating environmental sustainability, while enabling predictive maintenance and real-time monitoring to reduce costs and downtime. Feng et al. [109] introduced a data-driven AI framework that combines fuzzy clustering and TSVR with a metaheuristic optimizer, offering enhanced reservoir operation policies under uncertain conditions by identifying operational patterns and modeling complex nonlinear relationships. Similarly, Shaw et al. [110] leveraged ANNs as surrogate models to replace computationally intensive high-fidelity hydrodynamic simulations, enabling the optimization of hydropower generation while ensuring environmental compliance, particularly with dissolved oxygen constraints. In a different context, Uzlu et al. [111] demonstrated the efficacy of a hybrid ABC algorithm for predicting annual hydropower production in Turkey, achieving superior accuracy compared to traditional ANN methods. Furthermore, Hanoon et al. [112] applied multiple ML techniques, including ANN, ARIMA, and SVM, to forecast hydropower generation at the Three Gorges Dam, emphasizing AI’s adaptability in managing stochastic data across varying time scales—daily, monthly, and seasonal. Expanding on the theme of forecasting, Monteiro et al. [113] developed a short-term predictive model that integrates weather prediction tools with historical hydropower production data, significantly improving the accuracy of hourly generation forecasts for the Iberian electricity market. Fera and Spandonidis [114] introduced an IIoT-enabled AI framework to monitor and manage aging hydropower infrastructure, addressing challenges like corrosion and structural degradation. Villeneuve et al. [115] explored the application of AI in hydropower scheduling and optimization, utilizing ML to improve short-, medium-, and long-term energy production planning. Karakatsanis and Theodossiou [116] developed AI-based methods to integrate micro-hydropower generation into existing water distribution networks, optimizing energy recovery from pressure fluctuations. Additionally, Hajimohammadali et al. [117] proposed hybrid DL models for predictive maintenance, which achieved early fault detection and improved operational efficiency in hydropower plants. Beyond optimization, Siniosoglou et al. [118] addressed cybersecurity challenges in hydropower by proposing a DL-based anomaly detection system utilizing GANs and autoencoders to identify operational anomalies and detect cyberattacks effectively. Finally, Cui et al. [119] conducted a comprehensive investigation into techniques for detecting false data injection attacks, with a particular emphasis on their application in smart grids. The study underscores the critical role of AI-driven detection mechanisms in ensuring the integrity of operational data and mitigating cybersecurity risks. Furthermore, it provides valuable insights and strategic guidance for enhancing the resilience of hydropower systems against malicious attacks. However, as AI continues to drive advancements in smart grid integration and hybrid energy systems, future prospects must address emerging cybersecurity threats to ensure resilient and secure energy infrastructures.

5.3. Nuclear Power Plants

Similar to its applications in hydropower plants, AI holds immense potential for enhancing the efficiency and safety of nuclear power plants. The integration of AI strategies into control systems, which have been extensively utilized in nuclear power plants, can significantly improve operational efficiency, minimize downtime, and enhance monitoring capabilities to ensure safe and stable operations. Liu et al. introduced a ML-based prediction model for autonomous control of small reactors, using SVR and particle filtering to enhance system performance and reduce operator intervention [120]. Saeed et al. developed a Core Monitoring System leveraging an ANN for real-time surveillance of reactor power and safety parameters, achieving significant improvements in monitoring efficiency and accuracy [121]. Koo et al. demonstrated the use of DNN for predicting reactor vessel water levels during severe accidents, showcasing the potential of AI in addressing critical instrumentation challenges [122]. These advancements highlight AI’s transformative role in optimizing reactor control, predicting system behavior, and boosting operational efficiency.
Ensuring safety remains a paramount concern in nuclear power plant operations, and AI has played a crucial role in addressing these challenges. By combining knowledge-driven and data-driven approaches, AI has enhanced fault detection in nuclear systems, improving accuracy [123]. In addition to fault detection, Fu et al. proposed an interpretable AI framework using GRU models and SHAP analysis to predict severe accidents, enabling more effective emergency responses [124]. To strengthen cybersecurity, Lee and Huh applied big data analysis and AI to detect and mitigate APT in nuclear facilities [125]. Similarly, Yockey et al. assessed the cybersecurity risks of ML-driven autonomous control systems, stressing the need for robust safeguards against vulnerabilities in digital twin technologies [126]. Kumari et al. further highlighted AI’s role in behavioral analysis and anomaly detection to protect nuclear plants from cyberattacks, including zero-day exploits [127]. Jorge E. Núñez Mc Leod and Selva S. Rivera advanced computational intelligence in reliability engineering, applying probabilistic analysis and decision diagram methods to optimize safety and availability in multi-state and complex nuclear power systems [128]. Almoqbil et al. introduced FusionGuard, a ML-based framework for nuclear power plants that combines supervised and unsupervised models, real-time data fusion, and explainable AI to detect and mitigate ransomware and spyware, enhancing cybersecurity and operational resilience [129]. Together, these studies emphasize AI’s critical role in enhancing safety through improved predictive capabilities, anomaly detection, and proactive cybersecurity measures, ensuring the resilience of nuclear power plants against both operational and cyber threats.

5.4. Hydrogen Energy Systems

AI-driven strategies have enabled significant advancements in hydrogen energy systems, enhancing production efficiency, improving predictive modeling accuracy, optimizing operational parameters, and facilitating integration into hybrid energy networks. These advancements collectively contribute to the increased reliability and sustainability of hydrogen energy systems. For instance, Ahmed et al. introduced the OHGR model, using DL and SHAP analysis to predict hydrogen production in PEM fuel cells with high accuracy, while optimizing critical parameters like temperature and pressure to improve sustainability [130]. Similarly, Devasahayam demonstrated the effectiveness of DL models in predicting hydrogen production during co-gasification processes, highlighting the importance of feedstock particle size in enhancing hydrogen yields [131]. Zhang et al. applied advanced AI optimization algorithms, such as global dynamic harmony search, to optimize off-grid hybrid wind–hydrogen systems, reducing costs and improving system reliability and robustness [132]. In hybrid energy systems, Hwangbo et al. used DL-based supply–demand forecasting to manage renewable electricity and hydrogen production in self-sustaining energy networks, showcasing the flexibility and sustainability of hydrogen as an energy carrier [133]. These studies emphasize the transformative impact of AI in optimizing hydrogen energy systems, driving innovation, and improving efficiency.
Safety remains a critical focus in hydrogen energy systems, and AI has been pivotal in mitigating risks and ensuring secure operations. Zhao et al. developed a strategic cyberattack simulation framework that integrates CNN and double deep Q-networks to identify vulnerabilities in hybrid hydrogen power networks, highlighting the importance of proactive cybersecurity measures [134]. Halgamuge reviewed the role of DL in securing renewable energy supply chains, emphasizing its potential to detect cyber threats and prevent disruptions in hydrogen-based systems [135]. Sawas et al. proposed ML-based detection schemes to counter false data injection attacks on integrated PtG and GfG systems, demonstrating the significance of real-time monitoring and anomaly detection in ensuring operational resilience [136]. Together, these studies underscore AI’s critical role in optimizing hydrogen energy systems, enhancing safety, and supporting the secure and sustainable development of hydrogen technologies.

5.5. Geothermal Energy Systems and Biomass Energy Systems

For geothermal energy, AI has been instrumental in improving the efficiency of resource exploration and system management, leveraging ML and hybrid DL models to map subsurface temperatures and optimize energy extraction. For instance, Lialestani et al. employed a hybrid adaptive multitask DL framework to generate 3D geothermal maps, integrating modified firefly algorithms to identify geothermal hotspots with high accuracy [137]. Moraga et al. utilized AI and ML to analyze geophysical data, enabling better geothermal exploration through automated surface anomaly detection and prediction of geothermal potential [138].
AI-driven models are increasingly being utilized to predict biofuel yields, optimize conversion processes, and efficiently manage supply chains. Looking ahead, AI holds the potential to address key challenges in biomass energy systems, including improving feedstock utilization, reducing greenhouse gas emissions, and enhancing scalability, ultimately paving the way for a more sustainable energy future. Meena et al. demonstrated the application of AI in biochemical and thermochemical conversion technologies, using ML to optimize critical parameters such as temperature, pressure, and feedstock composition to maximize biofuel yields [139]. Similarly, Okolie et al. emphasized AI’s role in accelerating the discovery of new catalytic materials and microorganisms for biofuel production, enabling faster and more accurate predictions of reaction pathways and outcomes [140].

6. Conclusions

6.1. Summary of Research Achievements and Challenges

The integration of AI into renewable energy systems has emerged as a critical pathway towards achieving sustainable and efficient energy solutions. This review has comprehensively explored the applications and security challenges of AI in various renewable energy systems, including wind power, photovoltaic systems, and energy storage systems. The findings highlight significant advancements in prediction accuracy, system performance optimization, and fault detection and diagnosis capabilities. AI-driven models, particularly LSTM and hybrid models, have significantly improved short-term wind power prediction accuracy, enabling better grid integration and operational efficiency. DL techniques, such as CNN and LSTM, have enhanced the precision of PV power output predictions, even under varying weather conditions, and have been instrumental in improving system reliability and maintenance. AI has also played a crucial role in optimizing charging and discharging strategies, battery health management, and overall system scheduling for energy storage systems, enhancing their efficiency and economic viability. However, the widespread adoption of AI has introduced new security challenges, including data privacy, system stability, and algorithmic vulnerabilities. Advanced encryption, secure data transmission protocols, and robust access control mechanisms are essential to mitigate these risks.
The application of AI in renewable energy systems not only optimizes the performance of individual components but also enhances the overall stability and reliability of the energy grid. By improving prediction accuracy and enabling real-time monitoring and control, AI facilitates the integration of renewable energy sources into the existing power infrastructure, reducing the reliance on fossil fuels and mitigating the impacts of climate change. Furthermore, AI-driven fault detection and diagnosis systems reduce downtime and maintenance costs, ensuring the continuous operation of renewable energy systems.

6.2. Comparison with Prior Reviews and Future Research

When comparing the findings of this review with prior reviews, some similarities and differences emerge. Prior reviews also recognized the potential of AI in enhancing renewable energy systems, such as improving power generation efficiency and prediction accuracy. However, many earlier reviews often focused on a single aspect of AI applications in renewable energy, like only discussing AI in wind power or PV systems, rather than providing a comprehensive overview of multiple renewable energy systems as this review does. Regarding security challenges, while previous reviews touched upon data security and privacy issues, they did not delve as deeply into the specific algorithm security issues in the context of renewable energy systems, especially in relation to the unique operational characteristics of energy storage systems and other emerging renewable energy technologies. This review, on the other hand, has provided a more in-depth and systematic analysis of security challenges across different renewable energy systems, which is a significant contribution compared to prior reviews. Additionally, in terms of future research directions, previous reviews mainly emphasized the development of more accurate prediction models. This review, while acknowledging the importance of prediction, also highlights the need for more sophisticated and adaptive AI models, the enhancement of fault detection and diagnosis systems, and the establishment of comprehensive cybersecurity frameworks and relevant policies, offering a broader and more forward-looking perspective for future research in this field.
Future research should focus on developing more sophisticated and adaptive AI models that can handle the increasing complexity and variability of renewable energy systems. This includes the integration of multi-source data and the development of hybrid models combining multiple AI techniques. Additionally, the enhancement of AI-driven fault detection and diagnosis systems, the development of robust cybersecurity frameworks, and the establishment of policies and regulations to ensure the safe and ethical use of AI in the energy sector are critical areas for future exploration.
In conclusion, the integration of AI into renewable energy systems represents a significant step forward in the quest for sustainable and efficient energy solutions. While the technology offers substantial benefits, it also presents new challenges that require concerted efforts from researchers, industry practitioners, and policymakers. By addressing these challenges and leveraging the full potential of AI, we can pave the way for a more sustainable and secure energy future.

Funding

This work was supported in part by the State Grid Information and Telecommunication Group Co., Ltd., which coordinates scientific and technological projects (SGIT0000XTJS2401078).

Conflicts of Interest

Authors Hui Xiang, Xiaolei Li, Xiao Liao and Wei Cui were employed by the State Grid Information and Telecommunication Group Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Abbreviations

The following abbreviations are used in this manuscript:
AIArtificial Intelligence
PVPhotovoltaic
MLMachine Learning
NWPNumerical Weather Prediction
NARXNonlinear Autoregressive Exogenous
LSTMLong Short-Term Memory Network
CNN-BiLSTMConvolutional Neural Network-Bidirectional Long Short-Term Memory
EMDEmpirical Mode Decomposition
SVMSupport Vector Machine
RFRandom Forest
PCAPrincipal Component Analysis
DLDeep Learning
ANNArtificial Neural Network
ELMExtreme Learning Machine
RNNRecurrent Neural Networks
IIVIntra-Interval Variation
GATGraph Attention Networks
IGWOImproved Grey Wolf Algorithm
ESNEcho State Network
WTWavelet Transforms
XAIExplainable Artificial Intelligence
VRVirtual Reality
ARAugmented Reality
SOCState of Charge
SOHState of Health
DQNDeep Q Networks
TSVRTwin Support Vector Regression
ABCArtificial Bee Colony
IIoTInternet of Things
GANsGenerative Adversarial Networks
SVRSupport Vector Regression
DNNDeep Neural Networks
APTAdvanced Persistent Threats
OHGROptimized Hydrogen Generation-based Regression
PEMProton Exchange Membrane
PtGPower-to-Gas
GfGGas-fired Generation

References

  1. Qin, C.; Yu, Y. Security region based probabilistic small signal stability analysis for power systems with wind power integration. Autom. Electr. Power Syst. 2014, 38, 43–48. [Google Scholar]
  2. Zhu, Q.; Li, J.; Qiao, J.; Shi, M.; Wang, C. Application and Prospect of Artificial Intelligence Technology in Renewable Energy Forecasting. Proc. Chin. Soc. Electr. Eng. 2023, 43, 3027–3047. [Google Scholar] [CrossRef]
  3. Pan, G.; Zhang, H.; Ju, W.; Yang, W.; Qin, C.; Pei, L.; Sun, Y.; Wang, R. A prediction method for ultra short-term wind power prediction basing on long short-term memory network and extreme learning machine. In Proceedings of the 2020 Chinese Automation Congress (CAC), Shanghai, China, 6–8 November 2020; pp. 7608–7612. [Google Scholar]
  4. Liu, X.; Li, X. Error Compensation-Considered Wind Power Forecasting: A Hybrid Deep Learning Method. In Proceedings of the 2023 China Automation Congress (CAC), Shanghai, China, 6–8 November 2023; pp. 7319–7324. [Google Scholar]
  5. Zhang, B.; Liu, H.; Zhuang, G.; Liu, L.; Wu, W. Data-driven wind farm Volt/Var control based on deep reinforcement learning. In Proceedings of the 2020 IEEE 4th Conference on Energy Internet and Energy System Integration (EI2), Wuhan, China, 30 October–1 November 2020; pp. 2758–2763. [Google Scholar]
  6. Ahmadi, A.; Talaei, M.; Sadipour, M.; Amani, A.M.; Jalili, M. Deep federated learning-based privacy-preserving wind power forecasting. IEEE Access 2022, 11, 39521–39530. [Google Scholar] [CrossRef]
  7. Al-Yahyai, S.; Charabi, Y.; Gastli, A. Review of the use of numerical weather prediction (NWP) models for wind energy assessment. Renew. Sustain. Energy Rev. 2010, 14, 3192–3198. [Google Scholar] [CrossRef]
  8. Foley, A.M.; Leahy, P.G.; Marvuglia, A.; McKeogh, E.J. Current methods and advances in forecasting of wind power generation. Renew. Energy 2012, 37, 1–8. [Google Scholar] [CrossRef]
  9. Bhavsar, S.; Pitchumani, R.; Ortega-Vazquez, M. Machine learning enabled reduced-order scenario generation for stochastic analysis of solar power forecasts. Appl. Energy 2021, 293, 116964. [Google Scholar] [CrossRef]
  10. Xu, Z.; Zhang, X. Short-term wind power prediction of wind farms based on LSTM + NARX neural network. In Proceedings of the 2021 International Conference on Computer Engineering and Application (ICCEA), Kunming, China, 25–27 June 2021; pp. 137–141. [Google Scholar]
  11. Liu, S.; Zhou, Z.; Zhao, H. Short-term wind power load forecasting based on ISSA-CNN-BiLSTM. In Proceedings of the 2024 3rd International Conference on Energy, Power and Electrical Technology (ICEPET), Chengdu, China, 17–19 May 2024; pp. 1334–1339. [Google Scholar]
  12. Wang, Z.; Zhang, Z.; Wang, K.; Chu, L. Wind power prediction model based on long and short-term memory neural network. In Proceedings of the 2023 IEEE 3rd International Conference on Electronic Technology, Communication and Information (ICETCI), Changchun, China, 26–28 May 2023; pp. 754–758. [Google Scholar]
  13. van Heerden, L.; Vermeulen, H.; van Staden, C. Wind power forecasting using hybrid recurrent neural networks with empirical mode decomposition. In Proceedings of the 2022 IEEE International Conference on Environment and Electrical Engineering and 2022 IEEE Industrial and Commercial Power Systems Europe (EEEIC/I&CPS Europe), Prague, Czech Republic, 28 June–1 July 2022; pp. 1–6. [Google Scholar]
  14. Abedinia, O.; Lotfi, M.; Bagheri, M.; Sobhani, B.; Shafie-Khah, M.; Catalão, J.P. Improved EMD-based complex prediction model for wind power forecasting. IEEE Trans. Sustain. Energy 2020, 11, 2790–2802. [Google Scholar] [CrossRef]
  15. Wang, D.; Cui, X.; Niu, D. Wind power forecasting based on LSTM improved by EMD-PCA-RF. Sustainability 2022, 14, 7307. [Google Scholar] [CrossRef]
  16. Dong, H.; Xie, J.; Zhao, X. Wind farm control technologies: From classical control to reinforcement learning. Prog. Energy 2022, 4, 032006. [Google Scholar] [CrossRef]
  17. Tomin, N.; Kurbatsky, V.; Guliyev, H. Intelligent control of a wind turbine based on reinforcement learning. In Proceedings of the 2019 16th Conference on Electrical Machines, Drives and Power Systems (ELMA), Varna, Bulgaria, 6–8 June 2019; pp. 1–6. [Google Scholar]
  18. Soler, D.; Mariño, O.; Huergo, D.; de Frutos, M.; Ferrer, E. Reinforcement learning to maximize wind turbine energy generation. Expert Syst. Appl. 2024, 249, 123502. [Google Scholar] [CrossRef]
  19. Rasay, H.; Safaei, F.; Taghipour, S. A New Maintenance Plan for Wind Turbine Farms Using Reinforcement Learning. In Proceedings of the 2024 Annual Reliability and Maintainability Symposium (RAMS), Albuquerque, NM, USA, 22–25 January 2024; pp. 1–7. [Google Scholar]
  20. Chatterjee, J.; Dethlefs, N. Scientometric review of artificial intelligence for operations & maintenance of wind turbines: The past, present and future. Renew. Sustain. Energy Rev. 2021, 144, 111051. [Google Scholar]
  21. Zimroz, R.; Bartelmus, W.; Barszcz, T.; Urbanek, J. Statistical data processing for wind turbine generator bearing diagnostics. In Condition Monitoring of Machinery in Non-Stationary Operations: Proceedings of the Second International Conference “Condition Monitoring of Machinery in Non-Stationnary Operations” CMMNO’2012; Springer: Berlin/Heidelberg, Germany, 2012; pp. 509–518. [Google Scholar]
  22. Feng, Z.; Liang, M.; Zhang, Y.; Hou, S. Fault diagnosis for wind turbine planetary gearboxes via demodulation analysis based on ensemble empirical mode decomposition and energy separation. Renew. Energy 2012, 47, 112–126. [Google Scholar] [CrossRef]
  23. Leahy, K.; Hu, R.L.; Konstantakopoulos, I.C.; Spanos, C.J.; Agogino, A.M. Diagnosing wind turbine faults using machine learning techniques applied to operational data. In Proceedings of the 2016 IEEE International Conference on Prognostics and Health Management (ICPHM), Ottawa, ON, Canada, 20–22 June 2016; pp. 1–8. [Google Scholar]
  24. Si, Y.; Qian, L.; Mao, B.; Zhang, D. A data-driven approach for fault detection of offshore wind turbines using random forests. In Proceedings of the IECON 2017-43rd Annual Conference of the IEEE Industrial Electronics Society, Beijing, China, 29 October–1 November 2017; pp. 3149–3154. [Google Scholar]
  25. Clifton, A.; Kilcher, L.; Lundquist, J.; Fleming, P. Using machine learning to predict wind turbine power output. Environ. Res. Lett. 2013, 8, 024009. [Google Scholar] [CrossRef]
  26. Clifton, A.; Daniels, M.; Lehning, M. Effect of winds in a mountain pass on turbine performance. Wind Energy 2014, 17, 1543–1562. [Google Scholar] [CrossRef]
  27. Du, M.; Ma, S.; He, Q. A SCADA data based anomaly detection method for wind turbines. In Proceedings of the 2016 China International Conference on Electricity Distribution (CICED), Xi’an, China, 10–13 August 2016; pp. 1–6. [Google Scholar]
  28. Lu, Y.; Sun, L.; Zhang, X.; Feng, F.; Kang, J.; Fu, G. Condition based maintenance optimization for offshore wind turbine considering opportunities based on neural network approach. Appl. Ocean Res. 2018, 74, 69–79. [Google Scholar] [CrossRef]
  29. Qian, P.; Ma, X.; Wang, Y. Condition monitoring of wind turbines based on extreme learning machine. In Proceedings of the 2015 21st International Conference on Automation and Computing (ICAC), Glasgow, UK, 11–12 September 2015; pp. 1–6. [Google Scholar]
  30. Kulkarni, P.A.; Dhoble, A.S.; Padole, P.M. Deep neural network-based wind speed forecasting and fatigue analysis of a large composite wind turbine blade. Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. 2019, 233, 2794–2812. [Google Scholar] [CrossRef]
  31. Zhu, Q.; Li, H.; Wang, Z.; Chen, J.; Wang, B. Short-term wind power forecasting based on LSTM. Power Syst. Technol. 2017, 41, 3797–3802. [Google Scholar]
  32. Liu, Y.; Guan, L.; Hou, C.; Han, H.; Liu, Z.; Sun, Y.; Zheng, M. Wind power short-term prediction based on LSTM and discrete wavelet transform. Appl. Sci. 2019, 9, 1108. [Google Scholar] [CrossRef]
  33. Adlen, K.; Ridha, K. Recurrent neural network optimization for wind turbine condition prognosis. Diagnostyka 2022, 23, 2022301. [Google Scholar] [CrossRef]
  34. Gao, S.; Zeng, P.; Chen, Y.; Zheng, C.; Zhang, D.; Shang, R. A Dual Timescales Reactive Power Optimization Control Strategy Considering Voltage Security In Doubly-Fed Wind Farm. In Proceedings of the 2024 3rd International Conference on Energy, Power and Electrical Technology (ICEPET), Chengdu, China, 17–19 May 2024; pp. 1846–1850. [Google Scholar]
  35. Margaris, I.; Hansen, A.D.; Sørensen, P.; Hatziargyriou, N. Dynamic security issues in autonomous power systems with increasing wind power penetration. Electr. Power Syst. Res. 2011, 81, 880–887. [Google Scholar] [CrossRef]
  36. Ding, X.; Gong, Y.; Wang, C.; Zheng, Z. Artificial intelligence based abnormal detection system and method for wind power equipment. Int. J. Thermofluids 2024, 21, 100569. [Google Scholar] [CrossRef]
  37. Xu, Y.; Dong, Z.Y.; Xu, Z.; Meng, K.; Wong, K.P. An intelligent dynamic security assessment framework for power systems with wind power. IEEE Trans. Ind. Inform. 2012, 8, 995–1003. [Google Scholar] [CrossRef]
  38. Zhang, Y.; Dong, Z.Y.; Xu, Y.; Su, X.; Fu, Y. Impact analysis of intra-interval variation on dynamic security assessment of wind-energy power systems. In Proceedings of the 2020 IEEE Power & Energy Society General Meeting (PESGM), Montreal, QC, Canada, 2–6 August 2020; pp. 1–5. [Google Scholar]
  39. Shu, Y.; Yuan, Q.; Ke, W.; Kou, L. Security Access Control Method for Wind-Power-Monitoring System Based on Agile Authentication Mechanism. Electronics 2022, 11, 3938. [Google Scholar] [CrossRef]
  40. Zhao, R.; Chen, J.; Hou, Z.; Li, B.; Lin, M.; Duan, M. A security early warning method of power grid based on failure risk assessment. In Proceedings of the 2021 IEEE Sustainable Power and Energy Conference (iSPEC), Nanjing, China, 23–25 December 2021; pp. 1627–1633. [Google Scholar]
  41. Taghezouit, B.; Harrou, F.; Sun, Y.; Merrouche, W. Model-based fault detection in photovoltaic systems: A comprehensive review and avenues for enhancement. Results Eng. 2024, 21, 101835. [Google Scholar] [CrossRef]
  42. Lukas, F.; Viktor, U.; Claudio, R.; Bernhard, G.; Manuel, G. Fault detective: Automatic fault-detection for solar thermal systems based on artificial intelligence. Sol. Energy Adv. 2023, 3, 100033. [Google Scholar]
  43. Jia, Y.; Chen, G.; Zhao, L. Defect detection of photovoltaic modules based on improved VarifocalNet. Sci. Rep. 2024, 14, 15170. [Google Scholar] [CrossRef]
  44. Christian, U.; Christian, M.; Johannes, S.; Rutger, S.; Carolin, U. Explainable artificial intelligence for photovoltaic fault detection: A comparison of instruments. Sol. Energy 2023, 249, 139–151. [Google Scholar]
  45. Zhounan, W.; Peter, Z.; Basaran Bahadir, K.; Mirko, K. Drone-Based Solar Cell Inspection with Autonomous Deep Learning. In Infrastructure Robotics: Methodologies, Robotic Systems and Applications; IEEE: New York, NY, USA, 2024; pp. 337–365. [Google Scholar]
  46. Polymeropoulos, I.; Bezyrgiannidis, S.; Vrochidou, E.; Papakostas, G.A. Enhancing Solar Plant Efficiency: A Review of Vision-Based Monitoring and Fault Detection Techniques. Technologies 2024, 12, 175. [Google Scholar] [CrossRef]
  47. Mansouri, N.; Zitouni, N.; Mouelhi, A. AI Innovations in Photovoltaic Power Prediction. In Proceedings of the 2024 IEEE International Conference on Artificial Intelligence & Green Energy (ICAIGE), Yasmine Hammamet, Tunisia, 10–12 October 2024; pp. 1–6. [Google Scholar]
  48. Salman, D.; Direkoglu, C.; Kusaf, M.; Fahrioglu, M. Hybrid deep learning models for time series forecasting of solar power. Neural Comput. Appl. 2024, 36, 9095–9112. [Google Scholar] [CrossRef]
  49. Bakht, M.P.; Mohd, M.N.H.; Ibrahim, B.S.K.S.M.K.; Khan, N.; Sheikh, U.U.; Ab Rahman, A.A.-H. Advanced automated machine learning framework for photovoltaic power output prediction using environmental parameters and SHAP interpretability. Results Eng. 2025, 25, 103838. [Google Scholar] [CrossRef]
  50. Shao, X.; Pu, C.; Zhang, Y.; Kim, C.S. Domain Fusion CNN-LSTM for Short-Term Power Consumption Forecasting. IEEE Access 2020, 8, 188352–188362. [Google Scholar] [CrossRef]
  51. Xiang, X.; Li, X.; Zhang, Y.; Hu, J. A short-term forecasting method for photovoltaic power generation based on the TCN-ECANet-GRU hybrid model. Sci. Rep. 2024, 14, 6744. [Google Scholar] [CrossRef]
  52. Zhang, S.; Zhang, S.; Yu, J.J.Q.; Wei, X. ST-AGNet: Dynamic power system state prediction with spatial–temporal attention graph-based network. Appl. Energy 2024, 365, 123252. [Google Scholar] [CrossRef]
  53. Chen, R.; Cao, J.; Zhang, D. Probabilistic Prediction of Photovoltaic Power Using Bayesian Neural Network—LSTM Model. In Proceedings of the 2021 IEEE 4th International Conference on Renewable Energy and Power Engineering (REPE), Beijing, China, 9–11 October 2021; pp. 294–299. [Google Scholar]
  54. Amadou, B.; Alphousseyni, N.; Mbaye, N.E.h.; Senghane, M. Power optimization of a photovoltaic system with artificial intelligence algorithms over two seasons in tropical area. MethodsX 2023, 10, 101959. [Google Scholar]
  55. Ge, W.; Wang, X. PSO–LSTM–Markov Coupled Photovoltaic Power Prediction Based on Sunny, Cloudy and Rainy Weather. J. Electr. Eng. Technol. 2024, 20, 935–945. [Google Scholar] [CrossRef]
  56. Feng, N.; Song, D.; Liu, Z.; Wu, G. A Novel LSTM-XGBoost Model Optimized by SSA for Predicting Short-Term Photovoltaic Power. In Proceedings of the 2023 3rd International Conference on Energy, Power and Electrical Engineering (EPEE), Wuhan, China, 15–17 September 2023; pp. 128–133. [Google Scholar]
  57. Al, S.T.; Ahmed, H.; Gaeid, K.S.; Adnan, A.-S.; Yaseen, A.-H.; Smadi, K.A. Artificial intelligent control of energy management PV system. Results Control Optim. 2024, 14, 100343. [Google Scholar]
  58. ur Rehman, A.; Muhammad, A.; Sheeraz, I.; Aqib, S.; Nasim, U.; Al, O.S. Artificial Intelligence-Based Control and Coordination of Multiple PV Inverters for Reactive Power/Voltage Control of Power Distribution Networks. Energies 2022, 15, 6297. [Google Scholar] [CrossRef]
  59. Wang, Y.; Qi, Y.; Li, J.; Huan, L.; Li, Y.; Xie, B.; Wang, Y. The Wind and Photovoltaic Power Forecasting Method Based on Digital Twins. Appl. Sci. 2023, 13, 8374. [Google Scholar] [CrossRef]
  60. Zhang, S.; Wang, J.; Liu, H.; Tong, J.; Sun, Z. Prediction of energy photovoltaic power generation based on artificial intelligence algorithm. Neural Comput. Appl. 2020, 33, 821–835. [Google Scholar] [CrossRef]
  61. Mandal, P.; Madhira, S.T.S.; Haque, A.U.; Meng, J.; Pineda, R.L. Forecasting Power Output of Solar Photovoltaic System Using Wavelet Transform and Artificial Intelligence Techniques. Procedia Comput. Sci. 2012, 12, 332–337. [Google Scholar] [CrossRef]
  62. Salazar-Peña, N.; Tabares, A.; González-Mancera, A. AI-Powered Dynamic Fault Detection and Performance Assessment in Photovoltaic Systems. arXiv 2024, arXiv:2409.00052. [Google Scholar]
  63. Kim, J.; Ibrahim, H.; Wang, S.; Mete, A.; Xie, L.; Enjeti, P.; Kumar, P.R. Cyber-Secure and Safe Operation of Solar Photovoltaic Power Distribution Systems. In Proceedings of the 2024 IEEE Applied Power Electronics Conference and Exposition (APEC), Long Beach, CA, USA, 25–29 February 2024; pp. 1280–1287. [Google Scholar]
  64. Murat, K.; Umit, C.; Vinayak, S.; Ozgur, G. Gaining Insight Into Solar Photovoltaic Power Generation Forecasting Utilizing Explainable Artificial Intelligence Tools. IEEE Access 2020, 8, 187814–187823. [Google Scholar]
  65. Nozarian, M.; Fereidunian, A.; Hajizadeh, A.; Shahinzadeh, H. Exploring Social Capital in Situation-Aware and Energy Hub-Based Smart Cities: Towards a Pandemic-Resilient City. Energies 2023, 16, 6479. [Google Scholar] [CrossRef]
  66. Shaneh, M.; Shahinzadeh, H.; Iranpour, M.; Bayindir, R.; Nafisi, H.; Marzband, M. Risk curtailment assessment in smart deregulated grid with the presence of renewable and storage sources. In Proceedings of the 2020 6th Iranian Conference on Signal Processing and Intelligent Systems (ICSPIS), Mashhad, Iran, 23–24 December 2020; pp. 1–8. [Google Scholar]
  67. Karimi, A.; Aminifar, F.; Fereidunian, A.; Lesani, H. Energy storage allocation in wind integrated distribution networks: An MILP-Based approach. Renew. Energy 2019, 134, 1042–1055. [Google Scholar] [CrossRef]
  68. Nadeem, F.; Hussain, S.M.S.; Tiwari, P.K.; Goswami, A.K.; Ustun, T.S. Comparative Review of Energy Storage Systems, Their Roles and Impacts on Future Power Systems. IEEE Access 2018, 7, 4555–4585. [Google Scholar] [CrossRef]
  69. Zhou, Y.; Wang, X.; Mu, X.; Long, Z.; Zhou, L. Energy Storage Techniques Applied in Smart Grid. In Communications, Signal Processing, and Systems; Springer: Singapore, 2020. [Google Scholar]
  70. Karanki, S.B.; Xu, D.; Venkatesh, B.; Singh, B.N. Optimal location of battery energy storage systems in power distribution network for integrating renewable energy sources. In Proceedings of the 2013 IEEE Energy Conversion Congress and Exposition, Denver, CO, USA, 15–19 September 2013; pp. 4553–4558. [Google Scholar]
  71. Kolokotsa, D.; Kampelis, N.; Mavrigiannaki, A.; Gentilozzi, M.; Paredes, F.; Montagnino, F.; Venezia, L. On the integration of the energy storage in smart grids: Technologies and applications. Energy Storage 2019, 1, e50. [Google Scholar] [CrossRef]
  72. Biswal, B.; Deb, S.; Datta, S.; Ustun, T.S.; Cali, U. Review on smart grid load forecasting for smart energy management using machine learning and deep learning techniques. Energy Rep. 2024, 12, 3654–3670. [Google Scholar] [CrossRef]
  73. Murnane, M.; Ghazel, A. A closer look at state of charge (SOC) and state of health (SOH) estimation techniques for batteries. Analog Devices 2017, 2, 426–436. [Google Scholar]
  74. Ho, W.S.; Macchietto, S.; Lim, J.S.; Hashim, H.; Muis, Z.A.; Liu, W.H. Optimal scheduling of energy storage for renewable energy distributed energy generation system. Renew. Sustain. Energy Rev. 2016, 58, 1100–1107. [Google Scholar] [CrossRef]
  75. Hu, Y.; Kuang, W.; Qin, Z.; Li, K.; Zhang, J.; Gao, Y.; Li, W.; Li, K. Artificial intelligence security: Threats and countermeasures. ACM Comput. Surv. (CSUR) 2021, 55, 1–36. [Google Scholar] [CrossRef]
  76. Dey, B.; Roy, B.; Datta, S.; Ustun, T. Forecasting ethanol demand in India to meet future blending targets: A comparison of ARIMA and various regression models. Energy Rep. 2023, 9, 411–418. [Google Scholar] [CrossRef]
  77. Das, R.P.; Samal, T.K.; Luhach, A.K. An Energy Efficient Evolutionary Approach for Smart City-Based IoT Applications. Math. Probl. Eng. 2023, 2023, 9937949. [Google Scholar] [CrossRef]
  78. Xin, Q.; Alazab, M.; Díaz, V.G.; Montenegro-Marin, C.E.; Crespo, R.G. A deep learning architecture for power management in smart cities. Energy Rep. 2022, 8, 1568–1577. [Google Scholar] [CrossRef]
  79. Yao, Z.; Zhang, T.; Wang, Q.; Zhao, Y. Short-Term Power Load Forecasting of Integrated Energy System Based on Attention-CNN-DBILSTM. Math. Probl. Eng. 2022, 2022, 1075698. [Google Scholar] [CrossRef]
  80. Khan, N.; Shahid, Z.; Alam, M.M.; Bakar Sajak, A.A.; Mazliham, M.; Khan, T.A.; Ali Rizvi, S.S. Energy management systems using smart grids: An exhaustive parametric comprehensive analysis of existing trends, significance, opportunities, and challenges. Int. Trans. Electr. Energy Syst. 2022, 2022, 3358795. [Google Scholar] [CrossRef]
  81. Kong, X.; Li, C.; Wang, C.; Zhang, Y.; Zhang, J. Short-term electrical load forecasting based on error correction using dynamic mode decomposition. Appl. Energy 2020, 261, 114368. [Google Scholar] [CrossRef]
  82. Widodo, D.; Iksan, N.; Udayanti, E. Renewable energy power generation forecasting using deep learning method. IOP Conf. Ser. Earth Environ. Sci. 2021, 700, 012026. [Google Scholar] [CrossRef]
  83. Ma, Y.; Chen, X.; Wang, L.; Yang, J. Study on smart home energy management system based on artificial intelligence. J. Sens. 2021, 2021, 9101453. [Google Scholar] [CrossRef]
  84. Li, M.; Xu, W.; Zhang, S.; Liu, L.; Hussain, A.; Hu, E.; Zhang, J.; Mao, Z.; Chen, Z. State of Health Estimation and Battery Management: A Review of Health Indicators, Models and Machine Learning. Materials 2025, 18, 145. [Google Scholar] [CrossRef]
  85. Akbar, K.; Zou, Y.; Awais, Q.; Baig, M.J.A.; Jamil, M. A machine learning-based robust state of health (SOH) prediction model for electric vehicle batteries. Electronics 2022, 11, 1216. [Google Scholar] [CrossRef]
  86. Yu, Q.; Nie, Y.; Guo, S.; Li, J.; Zhang, C. Machine learning enables rapid state of health estimation of each cell within battery pack. Appl. Energy 2024, 375, 124165. [Google Scholar] [CrossRef]
  87. Vidal, C.; Malysz, P.; Kollmeyer, P.; Emadi, A. Machine learning applied to electrified vehicle battery state of charge and state of health estimation: State-of-the-art. IEEE Access 2020, 8, 52796–52814. [Google Scholar] [CrossRef]
  88. Talluri, G.; Lozito, G.M.; Grasso, F.; Iturrino Garcia, C.; Luchetta, A. Optimal battery energy storage system scheduling within renewable energy communities. Energies 2021, 14, 8480. [Google Scholar] [CrossRef]
  89. Mehrjerdi, H.; Hemmati, R. Modeling and optimal scheduling of battery energy storage systems in electric power distribution networks. J. Clean. Prod. 2019, 234, 810–821. [Google Scholar] [CrossRef]
  90. Lu, Y.; Xiang, Y.; Huang, Y.; Yu, B.; Weng, L.; Liu, J. Deep reinforcement learning based optimal scheduling of active distribution system considering distributed generation, energy storage and flexible load. Energy 2023, 271, 127087. [Google Scholar] [CrossRef]
  91. Xin, H.; Zhang, M.; Seuss, J.; Wang, Z.; Gan, D. A real-time power allocation algorithm and its communication optimization for geographically dispersed energy storage systems. IEEE Trans. Power Syst. 2013, 28, 4732–4741. [Google Scholar] [CrossRef]
  92. Zheng, Y.; Hill, D.J.; Dong, Z.Y. Multi-agent optimal allocation of energy storage systems in distribution systems. IEEE Trans. Sustain. Energy 2017, 8, 1715–1725. [Google Scholar] [CrossRef]
  93. Lagorse, J.; Paire, D.; Miraoui, A. A multi-agent system for energy management of distributed power sources. Renew. Energy 2010, 35, 174–182. [Google Scholar] [CrossRef]
  94. Carvalho, R.P.; Marchiori, C.F.N.; Brandell, D.; Araujo, C.M. Artificial intelligence driven in-silico discovery of novel organic lithium-ion battery cathodes. Energy Storage Mater. 2022, 44, 313–325. [Google Scholar] [CrossRef]
  95. Citroni, R.; Mangini, F.; Frezza, F. Efficient Integration of Ultra-low Power Techniques and Energy Harvesting in Self-Sufficient Devices: A Comprehensive Overview of Current Progress and Future Directions. Sensors 2024, 24, 4471. [Google Scholar] [CrossRef]
  96. Scharf, J.; Chouchane, M.; Finegan, D.P.; Lu, B.; Redquest, C.; Kim, M.-c.; Yao, W.; Franco, A.A.; Gostovic, D.; Liu, Z.; et al. Bridging nano- and microscale X-ray tomography for battery research by leveraging artificial intelligence. Nat. Nanotechnol. 2022, 17, 446–459. [Google Scholar] [CrossRef]
  97. Qiu, W.; Wang, Y.; Liu, J. Multiscale computations and artificial intelligent models of electrochemical performance in Li-ion battery materials. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2021, 12, e1592. [Google Scholar] [CrossRef]
  98. Li, X.; Zhang, X.; Lin, F.; Blaabjerg, F. Artificial-Intelligence-Based Design for Circuit Parameters of Power Converters. IEEE Trans. Ind. Electron. 2022, 69, 11144–11155. [Google Scholar] [CrossRef]
  99. Del Grosso, G.; Pichler, G.; Palamidessi, C.; Piantanida, P. Bounding information leakage in machine learning. Neurocomputing 2023, 534, 1–17. [Google Scholar] [CrossRef]
  100. Goldblum, M.; Tsipras, D.; Xie, C.; Chen, X.; Schwarzschild, A.; Song, D.; Mądry, A.; Li, B.; Goldstein, T. Dataset security for machine learning: Data poisoning, backdoor attacks, and defenses. IEEE Trans. Pattern Anal. Mach. Intell. 2022, 45, 1563–1580. [Google Scholar] [CrossRef]
  101. Kapoor, S.; Narayanan, A. Leakage and the reproducibility crisis in ML-based science. arXiv 2022, arXiv:2207.07048. [Google Scholar]
  102. Zhang, W.; Tople, S.; Ohrimenko, O. Leakage of dataset properties in Multi-Party machine learning. In Proceedings of the 30th USENIX Security Symposium (USENIX Security 21), Online, 11–13 August 2021; pp. 2687–2704. [Google Scholar]
  103. Tonni, S.M. Information Leakage in Machine Learning Models. Doctoral Dissertation, Macquarie University, Melbourne, Australian, 2020. [Google Scholar]
  104. Sasse, L.; Nicolaisen-Sobesky, E.; Dukart, J.; Eickhoff, S.B.; Götz, M.; Hamdan, S.; Komeyer, V.; Kulkarni, A.; Lahnakoski, J.; Love, B.C. On Leakage in Machine Learning Pipelines. arXiv 2023, arXiv:2311.04179. [Google Scholar]
  105. Bae, H.; Jang, J.; Jung, D.; Jang, H.; Ha, H.; Lee, H.; Yoon, S. Security and privacy issues in deep learning. arXiv 2018, arXiv:1807.11655. [Google Scholar]
  106. Alberti, M.; Pondenkandath, V.; Wursch, M.; Bouillon, M.; Seuret, M.; Ingold, R.; Liwicki, M. Are you tampering with my data? In Proceedings of the European Conference on Computer Vision (ECCV) Workshops, Munich, Germany, 8–14 September 2018. [Google Scholar]
  107. Borghetti, J.-S. How can Artificial Intelligence be defective? In Liability for Artificial Intelligence and the Internet of Things; Nomos Verlagsgesells chaft: Baden, Germany, 2019; pp. 63–76. [Google Scholar]
  108. Yampolskiy, R.V. Artificial Intelligence Safety and Security; CRC Press: Boca Raton, FL, USA, 2018. [Google Scholar]
  109. Feng, Z.-k.; Niu, W.-j.; Zhang, T.-h.; Wang, W.-c.; Yang, T. Deriving hydropower reservoir operation policy using data-driven artificial intelligence model based on pattern recognition and metaheuristic optimizer. J. Hydrol. 2023, 624, 129916. [Google Scholar] [CrossRef]
  110. Shaw, A.R.; Smith Sawyer, H.; LeBoeuf, E.J.; McDonald, M.P.; Hadjerioua, B. Hydropower optimization using artificial neural network surrogate models of a high-fidelity hydrodynamics and water quality model. Water Resour. Res. 2017, 53, 9444–9461. [Google Scholar] [CrossRef]
  111. Uzlu, E.; Akpınar, A.; Özturk, H.T.; Nacar, S.; Kankal, M. Estimates of hydroelectric generation using neural networks with the artificial bee colony algorithm for Turkey. Energy 2014, 69, 638–647. [Google Scholar] [CrossRef]
  112. Hanoon, M.S.; Ahmed, A.N.; Razzaq, A.; Oudah, A.Y.; Alkhayyat, A.; Huang, Y.F.; El-Shafie, A. Prediction of hydropower generation via machine learning algorithms at three Gorges Dam, China. Ain Shams Eng. J. 2023, 14, 101919. [Google Scholar] [CrossRef]
  113. Monteiro, C.; Ramirez-Rosado, I.J.; Fernandez-Jimenez, L.A. Short-term forecasting model for aggregated regional hydropower generation. Energy Convers. Manag. 2014, 88, 231–238. [Google Scholar] [CrossRef]
  114. Fera, F.T.; Spandonidis, C. An Artificial Intelligence and Industrial Internet of Things-Based Framework for Sustainable Hydropower Plant Operations. Smart Cities 2024, 7, 496–517. [Google Scholar] [CrossRef]
  115. Villeneuve, Y.; Séguin, S.; Chehri, A. Ai-based scheduling models, optimization, and prediction for hydropower generation: Opportunities, issues, and future directions. Energies 2023, 16, 3335. [Google Scholar] [CrossRef]
  116. Karakatsanis, D.; Theodossiou, N. Smart hydropower water distribution networks, use of artificial intelligence methods and metaheuristic algorithms to generate energy from existing water supply networks. Energies 2022, 15, 5166. [Google Scholar] [CrossRef]
  117. Hajimohammadali, F.; Crisostomi, E.; Tucci, M.; Fontana, N. Evaluating Deep Learning Networks Versus Hybrid Network for Smart Monitoring of Hydropower Plants. Energies 2024, 17, 5670. [Google Scholar] [CrossRef]
  118. Siniosoglou, I.; Radoglou-Grammatikis, P.; Efstathopoulos, G.; Fouliras, P.; Sarigiannidis, P. A unified deep learning anomaly detection and classification approach for smart grid environments. IEEE Trans. Netw. Serv. Manag. 2021, 18, 1137–1151. [Google Scholar] [CrossRef]
  119. Cui, L.; Qu, Y.; Gao, L.; Xie, G.; Yu, S. Detecting false data attacks using machine learning techniques in smart grid: A survey. J. Netw. Comput. Appl. 2020, 170, 102808. [Google Scholar] [CrossRef]
  120. Zeng, Y.; Liu, J.; Sun, K.; Hu, L.-w. Machine learning based system performance prediction model for reactor control. Ann. Nucl. Energy 2018, 113, 270–278. [Google Scholar] [CrossRef]
  121. Saeed, A.; Rashid, A. Development of core monitoring system for a nuclear power plant using artificial neural network technique. Ann. Nucl. Energy 2020, 144, 107513. [Google Scholar] [CrossRef]
  122. Do Koo, Y.; An, Y.J.; Kim, C.-H.; Na, M.G. Nuclear reactor vessel water level prediction during severe accidents using deep neural networks. Nuclear Eng. Technol. 2019, 51, 723–730. [Google Scholar] [CrossRef]
  123. Qi, B.; Liang, J.; Tong, J. Fault diagnosis techniques for nuclear power plants: A review from the artificial intelligence perspective. Energies 2023, 16, 1850. [Google Scholar] [CrossRef]
  124. Fu, Y.; Zhang, D.; Xiao, Y.; Wang, Z.; Zhou, H. An interpretable time series data prediction framework for severe accidents in nuclear power plants. Entropy 2023, 25, 1160. [Google Scholar] [CrossRef]
  125. Lee, S.; Huh, J.-H. An effective security measures for nuclear power plant using big data analysis approach. J. Supercomput. 2019, 75, 4267–4294. [Google Scholar] [CrossRef]
  126. Yockey, P.; Erickson, A.; Spirito, C. Cyber threat assessment of machine learning driven autonomous control systems of nuclear power plants. Prog. Nucl. Energy 2023, 166, 104960. [Google Scholar] [CrossRef]
  127. Kumari, A.; Dubey, R.; Sharma, I. ShNP: Shielding Nuclear Plants from Cyber Attacks Using Artificial Intelligence Techniques. In Proceedings of the 2023 Annual International Conference on Emerging Research Areas: International Conference on Intelligent Systems (AICERA/ICIS), Kanjirapally, India, 16–18 November 2023; pp. 1–6. [Google Scholar]
  128. Mc Leod, J.E.N.; Rivera, S.S. Reliability Optimization of New Generation Nuclear Power Plants Using Artificial Intelligence. In Reliability Engineering and Computational Intelligence for Complex Systems: Design, Analysis and Evaluation; Springer: Berlin/Heidelberg, Germany, 2023; pp. 159–173. [Google Scholar]
  129. Almoqbil, A.H.N. Anomaly detection for early ransomware and spyware warning in nuclear power plant systems based on FusionGuard. Int. J. Inf. Secur. 2024, 23, 2377–2394. [Google Scholar] [CrossRef]
  130. Ahmed, R.; Shehab, S.A.; Elzeki, O.M.; Darwish, A.; Hassanein, A.E. An explainable AI for green hydrogen production: A deep learning regression model. Int. J. Hydrogen Energy 2024, 83, 1226–1242. [Google Scholar] [CrossRef]
  131. Devasahayam, S. Deep learning models in Python for predicting hydrogen production: A comparative study. Energy 2023, 280, 128088. [Google Scholar] [CrossRef]
  132. Zhang, W.; Maleki, A.; Pourfayaz, F.; Shadloo, M.S. An artificial intelligence approach to optimization of an off-grid hybrid wind/hydrogen system. Int. J. Hydrogen Energy 2021, 46, 12725–12738. [Google Scholar] [CrossRef]
  133. Hwangbo, S.; Nam, K.; Heo, S.; Yoo, C. Hydrogen-based self-sustaining integrated renewable electricity network (HySIREN) using a supply-demand forecasting model and deep-learning algorithms. Energy Convers. Manag. 2019, 185, 353–367. [Google Scholar] [CrossRef]
  134. Zhao, A.P.; Zhang, Q.; Alhazmi, M.; Hu, P.J.-H.; Zhang, S.; Yan, X. AI for science: Covert cyberattacks on energy storage systems. J. Energy Storage 2024, 99, 112835. [Google Scholar] [CrossRef]
  135. Halgamuge, M.N. Leveraging Deep Learning to Strengthen the Cyber-Resilience of Renewable Energy Supply Chains: A Survey. IEEE Commun. Surv. Tutor. 2024, 26, 2146–2175. [Google Scholar] [CrossRef]
  136. Sawas, A.M.; Khani, H.; Farag, H.E. On the resiliency of power and gas integration resources against cyber attacks. IEEE Trans. Ind. Inform. 2020, 17, 3099–3110. [Google Scholar] [CrossRef]
  137. Mirfallah Lialestani, S.P.; Parcerisa, D.; Himi, M.; Abbaszadeh Shahri, A. Generating 3D geothermal maps in Catalonia, Spain using a hybrid adaptive multitask deep learning procedure. Energies 2022, 15, 4602. [Google Scholar] [CrossRef]
  138. Moraga, J.; Duzgun, H.; Cavur, M.; Soydan, H. The geothermal artificial intelligence for geothermal exploration. Renew. Energy 2022, 192, 134–149. [Google Scholar] [CrossRef]
  139. Meena, M.; Shubham, S.; Paritosh, K.; Pareek, N.; Vivekanand, V. Production of biofuels from biomass: Predicting the energy employing artificial intelligence modelling. Bioresour. Technol. 2021, 340, 125642. [Google Scholar] [CrossRef]
  140. Okolie, J.A. Introduction of Machine Learning and artificial intelligence in biofuel technology. Curr. Opin. Green Sustain. Chem. 2024, 47, 100928. [Google Scholar] [CrossRef]
Figure 1. Application scenarios, algorithm models, security issues, and corresponding solutions of AI in wind power systems.
Figure 1. Application scenarios, algorithm models, security issues, and corresponding solutions of AI in wind power systems.
Energies 18 01931 g001
Figure 2. Application scenarios, applied algorithms, and security issues of AI in PV systems.
Figure 2. Application scenarios, applied algorithms, and security issues of AI in PV systems.
Energies 18 01931 g002
Figure 3. Application scenarios, applied algorithm, and security issues of AI in energy storage systems.
Figure 3. Application scenarios, applied algorithm, and security issues of AI in energy storage systems.
Energies 18 01931 g003
Figure 4. Application scenarios, applied algorithm, and security issues of AI in other renewable energy systems.
Figure 4. Application scenarios, applied algorithm, and security issues of AI in other renewable energy systems.
Energies 18 01931 g004
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Xiang, H.; Li, X.; Liao, X.; Cui, W.; Liu, F.; Li, D. Artificial Intelligence in Renewable Energy Systems: Applications and Security Challenges. Energies 2025, 18, 1931. https://doi.org/10.3390/en18081931

AMA Style

Xiang H, Li X, Liao X, Cui W, Liu F, Li D. Artificial Intelligence in Renewable Energy Systems: Applications and Security Challenges. Energies. 2025; 18(8):1931. https://doi.org/10.3390/en18081931

Chicago/Turabian Style

Xiang, Hui, Xiaolei Li, Xiao Liao, Wei Cui, Fengkai Liu, and Donghe Li. 2025. "Artificial Intelligence in Renewable Energy Systems: Applications and Security Challenges" Energies 18, no. 8: 1931. https://doi.org/10.3390/en18081931

APA Style

Xiang, H., Li, X., Liao, X., Cui, W., Liu, F., & Li, D. (2025). Artificial Intelligence in Renewable Energy Systems: Applications and Security Challenges. Energies, 18(8), 1931. https://doi.org/10.3390/en18081931

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop