Engineering a resilient smart grid: Practical defense mechanisms and deployable framework against evolving cyber-threats

Novel smart grids applications for energy management and traceability in heritage cities: Case study for city of Cuenca, Ecuador

The worldwide effort to reach carbon peak and neutrality objectives alongside energy market expansion has sped up renewable energy integration, like wind and solar power. The shift towards renewable energy integration introduces substantial uncertainties in power system scheduling and control processes, which test the limits of existing theoretical methods. The advanced reasoning and data-processing capabilities of Large Language Models (LLMs), with particular reference to their ability to analyze multimodal data, provide transformative potential for managing and controlling smart grids. This review examines how LLMs can tackle modern power system challenges while confirming their fit with the power sector’s expanding dependency on Artificial Intelligence (AI) technologies. We assess the requirements of modern power systems for such AI-based solutions, while evaluating how LLMs shape grid management and exploring their enabling technologies, such as model architecture and training methods, along with necessary data. Our review investigates how multimodal LLM technology serves different smart grids’ functions, including generation, transmission, distribution, consumption, and equipment management, to exhibit its adaptable nature in strengthening grid resilience and efficiency. • This review explores the role of multimodal Large Language Models (LLMs) in smart grid management, showing how their ability to integrate and process different types of data, including sensor readings, text logs, weather forecasts, and equipment images, can significantly improve decision-making, fault diagnosis, and operational planning in power systems. • The study analyzes the architectural and training aspects of multimodal LLMs, including the use of pretrained modular encoders, efficient fine-tuning methods such as Low-Rank adaptation (LoRA), and specialized loss functions, highlighting how these techniques enable adaptation to the specific needs of smart grid applications without lengthy retraining. • Practical considerations for industrial implementation are examined, covering multimodal data collection and preprocessing, domain-specific knowledge integration, intelligent task decomposition, and system-level integration, illustrating how LLMs can be seamlessly integrated into power system operating environments. • The review highlights the potential of multimodal LLMs to improve the resilience of the power grid, optimize the integration of renewable energy, and support human-machine collaboration, while outlining future research directions, such as domain-specific base models, physics-based architectures, and human-in-the-loop feedback, in order to further improve reliability and interpretability in critical infrastructure applications.

The increasing demand for renewable energy sources necessitates advanced methods to ensure grid stability and operational efficiency. Smart grids present a viable solution by enabling the seamless integration of renewable energy systems, such as solar photovoltaic (PV) and other sources. However, the performance and reliability of such systems are critically dependent on the performance and functionality of power inverters (e.g., solar PV inverters). This paper proposes the application of several decision tree algorithms, which are traditional decision tree (TDT), iterative dichotomizer tree (IDT), C4.5 algorithm, and CART algorithm, for fault prediction and diagnostics in solar PV inverters. The training and prediction phases of the decision tree models employ both key inverter specifications, such as nominal operating ranges, abnormal readings, and fault conditions, alongside critical operational parameters, including voltage, current, temperature, conversion efficiency, power factor (PF), and total harmonic distortion (THD). Also, a cost-benefit analysis is made to consider the cost-effectiveness of predictive maintenance of solar PV inverters with the presence of significant economic benefits. A substantial cost advantage is shown, where 12,000 $/year in losses is saved with an investment of merely 4000 $/year in support of a predictive maintenance model with a first year ROI of 200%. Overall, the results of all decision tree models, implemented in MATLAB, confirm that both TDT algorithm and CART algorithm attain a high classification accuracy of about 95% with TDT possessing the shortest fault prediction time of about 50 ms. Markedly, the CART algorithm proves effective when it comes to dealing with an up to 90% missing values. Additionally, the analysis demonstrates a direct proportional relationship between voltage, current, temperature, and conversion efficiency, while revealing an inverse relationship between conversion efficiency and THD. The proposed algorithms for fault prediction hold significant potential for enhancing the reliability of solar PV systems integrated within smart grid frameworks.

A robust hybrid framework for load forecasting in smart grid

Deep spatiotemporal feature fusion using CNN-LSTM autoencoder for real-time, phase-aware anomaly identification and localization in high-voltage power transmission networks

Federated learning (FL), as a decentralized machine learning paradigm, emerges as a pivotal approach to addressing the ecological challenges posed by traditional IoT systems. While existing research extensively explores FL in smart cities and healthcare, its potential for fostering sustainable IoT practices remains underexplored. This review fills this gap by exploring how FL can help in lowering the amount of carbon footprint and energy use of centralized IoT infrastructures. In comprehensive analyses, this study highlights the integration of FL with green computing concepts, it’s usage in various fields, including environmental monitoring and smart grids, and how it can interact with blockchain technology. Across selected case studies, federated learning is reported to improve runtime- or compute-related efficiency and predictive performance in specific environmental sensing settings, and FL–blockchain designs in smart-city settings are reported to reduce latency under the studied simulation assumptions. Despite these advancements, challenges like data heterogeneity, resource limitations, and privacy concerns exist. Proposed solutions include lightweight FL models, secure aggregation protocols, and adaptive resource allocation strategies. This review underscores FL’s transformative role in achieving a sustainable IoT ecosystem and identifies future research directions for robust and scalable green IoT implementations. • Federated Learning reduces IoT carbon footprint and energy consumption significantly. • Blockchain integration enhances data-sharing security and reduces latency by 20–30 ms. • FL improves computational efficiency up to 7.3 times and accuracy by over 13.2%. • Data heterogeneity and limited resources hinder the integration of FL in green IoT. • Lightweight FL models and secure protocols enhance efficiency and scalability.

A linear programming-based matheuristic for reliable customer–feeder mapping in smart grids

• A Darts-based machine learning framework is applied for electricity demand forecasting. • Comparative analysis of ML models- TSMixer, TiDE, Random Forest, and FB Prophet using real SCADA data from Tetouan, Morocco. • Min Max, Max Absolute, and Standard scaling preprocessing techniques are used, and Performance is assessed using MAE, RMSE, and MAPE. • TiDE and TSMixer demonstrated better forecasting accuracy and stability. • The Darts framework enhances forecasting accuracy and lowers RMSE and MAE by up to 60 Accurate electricity demand forecasting is essential for reliable power system operation, load balancing, and energy management in rapidly urbanizing cities, where modern smart grids exhibit rapid load variability. This study presents a comprehensive comparative analysis of electricity demand forecasting models using the Darts time-series framework for Tetouan, Morocco, based on SCADA data collected at 10-minute intervals across three distribution zones in 2017. Weather variables and historical load features are incorporated to capture nonlinear consumption patterns. The proposed methodology incorporates consistent preprocessing, feature engineering, and strictly chronological train–test splitting to avoid temporal leakage. Six models, Linear Regression, Random Forest, XGBoost, Prophet, TiDE, and TSMixer, are evaluated at both 10-minute and hourly intervals under a consistent preprocessing pipeline (Min-Max, Max Absolute, and Standard scaling). Model performance is assessed using standard error metrics, including MAE, RMSE, and MAPE. Results show that TSMixer and TiDE excel in fine-grained, short-term forecasting, with TiDE achieving a MAPE of 1.92% for combined 10-minute forecasts. While Random Forest provides a strong balance between accuracy and computational efficiency at hourly resolution with a MAPE of 1.04% and an MAE of 224.2. Compared with prior Tetouan-focused studies, the proposed framework shows substantial error reductions under identical data resolution. This work contributes a transparent benchmarking framework and actionable insights for short-term electricity demand forecasting in smart grid applications.

Plastic waste-derived fuel in diesel engine for 4 kW power generation supporting smart grid stabilization

Optimizing and evaluating deep learning techniques for stealthy false data injection attacks on smart grids

Multi-task transformer with adaptive attention and uncertainty-aware recovery for FDI attacks defense in smart grids

With the increasing incorporation of Renewable Energy Sources (RESs) like solar Photovoltaic (PV) systems, maintaining frequency stability has turned out to be a significant challenge owing to decreased system inertia. Despite numerous developments in Load Frequency Control (LFC), existing solutions largely overlooked the issue of Under Frequency Load Shedding (UFLS) relay failure during rapid frequency decline, which led to widespread blackouts. To address this critical gap, a novel intelligent control framework integrating the Fuzzy Doubleton Parabolic Inference System (FDPIS) and the Proportional Quad-Alpine Integral Derivative (PQAID) controller for UFLS relay failure mitigation and enhanced LFC in Solar-PV systems is proposed. Primarily, the Direct Current (DC) power from the solar module is fed into the DC-DC boost converter and Maximum Power Smoothstep Point Tracking (MPSPT) algorithm. A capacitor bank failure is detected using FDPIS, and voltage stabilization is ensured through a Savitzky-Golay Dynamic Polynomial-Z Voltage Restorer (SGDP-ZVR). To predict electrical load demand accurately, a hybrid Deep Learning (DL) model, Deep Dualplus Softshrink Pan–Long Short Term Inverse Parzen Memory (2DSP-LSTIPM), is employed, delivering a high accuracy of 98.98 % with a Root Mean Squared Error (RMSE) of 0.002. When demand exceeds thresholds, transmission overload is mitigated using an Inductive Snubber Cubic Circuits–STATCOM (IS2C-STATCOM). The frequency deviation is identified via FDPIS, followed by the Rate Of Change Of Frequency (ROCOF) analysis. If a UFLS relay failure is detected, then the PQAID controller is activated to ensure stable operation. The proposed PQAID achieves a peak time of 1.91 ms, significantly outperforming traditional PID, PI, and PD controllers in transient and overshoot metrics. Simulation results on the HEDGW dataset assess the proposed approach’s robustness and low time complexity. The system demonstrates superior relay fault detection (fuzzification/defuzzification times of 452ms/463ms) and faster rule generation (597 ms) compared to conventional fuzzy systems. Overall, the proposed methodology provides a comprehensive, real-time, and scalable solution for enhancing frequency stability, relay fault mitigation, and load management in solar PV-based smart grids. • Integrates FDPIS and PQAID for real-time UFLS relay failure detection and mitigation in solar PV systems. • Proposes novel 2DSP-LSTIPM deep learning model achieving 98.98 % demand prediction accuracy with RMSE of 0.002. • Introduces SGDP-ZVR for voltage stabilization during capacitor bank faults using Savitzky-Golay filtering. • Deploys IS2C-STATCOM for efficient transmission overload control with fast reactive power regulation. • Enables seamless SCADA/EMS integration via OPC-UA protocol for smart grid compatibility and deployment.

Robust fault detection and uncertainty quantification in smart grids using graph neural networks

Deep Tasmanian Aquila quantized Shor Simon fractal network for intelligent energy theft detection in smart grids

Unified rigid factor based reliability assessment towards reinforcement of electrical substations

Adaptive Reinforcement Learning for Dynamic Optimization of Sustainable Smart Power Grids Under Uncertain Conditions

Adaptive machine learning framework for fair and resilient load shedding in smart power distribution networks

Certificateless privacy-preserving multidimensional data aggregation scheme with fault-tolerance for fog-based smart grids

A day ahead scheduling model of a smart hydrogen-based microgrid taking into account PV production and electrical load demand forecasting errors