Bishwajit Dey

Modern power grids are exposed to False Data Injection (FDI) attacks due to their increasing dependency on digital sensing and measurement infrastructure. The adversaries deliberately manipulate sensor data for compromising control, system monitoring, and overall grid stability. Undetected attacks trigger cascading failures and severe economic disruption and bring challenges for developing a robust, data-driven detection framework for safeguarding critical energy infrastructure. The present study undertakes this challenge by introducing a structured, coordinated, feature-group-based synthetic FDI-inspired perturbation framework. Five architecturally different attack detection models deployed are Multilayer Perceptron (MLP), a Pure LSTM Autoencoder, Graph Convolutional Network (GCN), Graph Attention Network (GAT), and Standard GraphSAGE with an identical 20-step pipeline. A common k-NN cosine-similarity graph (k = 10) is used for the graph-based models. A model-specific hyperparameter grid search is conducted, which comprises 24 configurations for MLP, GCN, GAT, and GraphSAGE and 8 configurations for the LSTM autoencoder, reflecting its distinct unsupervised architecture. A 10-seed robustness test is performed to ensure rigorous, reproducible, and fair cross-model comparison. Under single-run evaluation, the GraphSAGE architecture achieves the strongest overall detection performance (ROC - AUC 0.7277 and PR - AUC 0.4679), outperforming the LSTM Autoencoder (ROC - AUC 0.6540), GCN (ROC - AUC 0.6284), MLP (ROC - AUC 0.5959), and GAT (ROC - AUC 0.4726). In multi-seed robustness evaluation results the GAT architecture emerges as the most reliable and stable detector with an accuracy of 0.8189 ± 0.0123 over ten random initialisations. The findings of this research motivate reporting both single-run and multi-seed robustness metrics when evaluating FDI attack detection models intended for real-world grid deployment. The novelty of this work lies in providing a unified, leakage-free comparative framework spanning graph- and non-graph-based deep learning architectures under identical experimental conditions. As the perturbations are FDI-inspired rather than state-estimation-aware, it provides a realistic and extensible benchmark for future power-grid cybersecurity research.

The increasing penetration of renewable energy sources (RES) and plug-in hybrid electric vehicles (PHEVs) has introduced significant frequency instability in interconnected microgrid (MG) systems, necessitating adaptive and robust control strategies for reliable operation. This paper proposes a fuzzy-explainable neural network (FxNN)-based distributed fractional-order PID (FOPID) controller for active frequency regulation in interconnected microgrids. The control problem is formulated within a Lyapunov-based optimization framework, where system stability is ensured through a recursively updated energy function and differential learning dynamics of the explainable neural network. The proposed cascaded FxNN-FOPID controller is evaluated under varying load disturbances and intermittent renewable power. Simulation results demonstrate superior dynamic performance compared with conventional single-loop FxNN and PSO-GSA-based controllers. The proposed controller achieves a minimum settling time of 3.0 s, representing improvements of 14.3% and 30.2%, respectively. Furthermore, the integral absolute error (IAE) is reduced to 0.6292 × 10 ⁻³ , 0.5929 × 10 ⁻³ , and 0.6000 × 10 ⁻³ for MG1, MG2, and MG3, respectively, while the integral time absolute error (ITAE) is reduced by up to 93.6%. The controller also minimizes frequency oscillations with a peak magnitude of 0.0003 p.u. and achieves improved Integral of Squared Error (ISE) 0.1040 × 10 ⁻⁵ and Integral of Time-weighted Squared Error (ITSE) 0.9918 × 10 ⁻³ values. Owing to its adaptive gain tuning capability, the proposed controller maintains robust performance without requiring manual retuning under varying operating conditions. Comparative analysis confirms that the proposed cascaded FxNN-based distributed FOPID controller provides faster dynamic response, improved disturbance rejection, and superior low-frequency oscillation damping, making it an effective solution for reliable frequency regulation in renewable-energy-integrated microgrid systems. Furthermore, the proposed framework supports Sustainable Development Goals (SDG 7: Affordable and Clean Energy, SDG 9: Industry, Innovation and Infrastructure, and SDG 13: Climate Action) by facilitating resilient microgrid operation, enhancing renewable energy integration, and promoting low-carbon power systems.

Sustaining power quality is an utmost priority for energy distributors in a modern power system integrated with distributed generation and nonlinear loads. The power quality disturbances (PQDs) comprising of multiple PQD events creates complexities in accurate detection and classification. Therefore, this study presents an effective, resilient and automated hybrid classifier for detection and classification of complex power quality disturbances for single occurrence as well as combinations of double, triple, and quadruple occurrences. To develop an effective and cohesive classifier, it is vital to identify the applicable features from the disturbance signal that can enhance data efficiency while capturing the signal’s fundamental qualities. Consequently, a set of features is retrieved utilizing the Stockwell Transform (ST), and the significant features are identified by the Chi-Square Test (CST). These features are utilized to train the Long Short-Term Memory (LSTM) network. The method has been validated on the signals generated using Power System Computer Aided Design (PSCAD) and OPAL-Real Time (OPAL-RT) for the IEEE 9-bus and 33-bus systems. The novelty of proposed hybrid method lies in the real time validation under noisy conditions. Additionally, to signify the superiority of the suggested method, the performance of the classifier has been compared with a multi-layer feedforward neural network and existing approaches documented in the manuscript. The proposed hybrid method given 99.11% accuracy which is higher in comparison of existing methods. The results confirm the superiority of the suggested hybrid technique for the real-time detection and classification of PQD signals.

The economic and low-carbon functionality of a microgrid is considerably enhanced by Electric Vehicles (EVs) relative to Battery Energy Storage Systems (BESS) due to their adaptable demand-side engagement, bidirectional power transfer, and mobility. PHEVs use vehicle-to-grid (V2G) technology, facilitating real-time balancing of energy use, lowering peak loads, and supporting flexible power sharing. To find the most cost-effective approach, BESS and EVs are integrated into a low voltage microgrid system and compare the two load shifting strategies. By adjusting the number of EVs from 5 to 20, sensitivity analysis is used to explore the impact of critical factors on system performance, such as battery capacity, state of charge, and operating time. The results shed light on how reliable, economical, and practical the system is in terms of the optimal load shifting method as compared to the peak clipping one. The proposed method ensures cost-effective operation, enhances energy planning, and improves microgrid sustainability, thereby strengthening grid stability and promoting greater integration of renewable energy into modern grid frameworks.

The rapid integration of electric vehicle (EV) infrastructure, particularly EV charging stations (EVCS) and battery swapping stations (BSS), has introduced new challenges to modern distribution networks due to dynamic and uncertain load profiles. This paper proposes a comprehensive profit maximization framework aimed at improving economic outcomes for three key stakeholder groups: EVCS operators, BSS operators, and distribution network consumers. The model integrates demand-side management strategies such as load shifting and load curtailment to mitigate undesirable load fluctuations while enhancing operational efficiency. A hybrid metaheuristic optimization technique combining Differential Evolution with Prairie Dog Optimization is used to solve the multi-entity coordination problem. Extensive simulation studies are carried out on a 118-bus radial distribution network under six different case studies. Results indicate that simultaneous profit maximization across all stakeholders yields superior economic performance compared to individual optimization. The proposed framework achieves up to 18.7% higher total profit. Specifically, EVCS operators, BSS operators, and demand-side participants benefit from profit improvements of 21.9%, 24.2%, 49.9%, and 2.5%, respectively, highlighting the economic viability of the proposed coordinated approach across all six operational cases.