With grid modernization, smart inverters are increasingly used to execute advanced controls for distribution network reliability. However, this also increases the cyber-attack space. This paper focuses on the defense approaches to restore the system to normal operation circumstances in the presence of cyber-attacks. A unique deep reinforcement learning (DRL) method is developed to minimize voltage violations and reduce power losses for impacted feeders. The defense problem is reformulated as a Markov decision-making process to dynamically control DERs while minimizing load shedding. This is achieved via an improved soft actor-critic (SAC)-based DRL algorithm, which can govern DER set points and load-shedding scenarios in discrete and continuous modes via the auto-tune entropy and Gaussian policy features. Numerical comparison results on the modified IEEE 123-node system with other control approaches, such as Volt-VAR (VV), Volt-Watt (VW), and model predictive control (MPC) show that the proposed method can eliminate voltage violations and provide feasible control actions that perform complete mitigation of cyber-threats.

This letter demonstrates a proof-of-concept validation of the Large Language Model (LLM) for smart inverter cyberattack detection through textual control commands. The proposed method can detect the manipulation of Volt/VAR curves and the comparison results with state-of-the-art machine learning techniques highlight its efficacy in identifying cyber attacks. Test results obtained from a real distribution feeder in Colorado, USA, validate the high accuracy for attack classification as well as demonstrate the potential of LLMs in adding a robust security layer for cyber-attacks.

This paper proposes a safe reinforcement learning (RL)-based transient stability emergency control (TSEC) method for islanded microgrids. RL requires extensive interaction with the environment to learn control strategies, hence, a data-driven approach is used as a substitute for time-consuming time-domain simulation calculations. Deep sigma point processes (DSPP), which is a Gaussian process model, is utilized to predict the normal distribution of transient stability of microgrids and to construct a transient stability chance constraint. Reward-constrained policy optimization (RCPO) can simultaneously achieve objective prediction, policy learning, and constraint cost coefficient update across multiple timescales. RCPO interacts with the DSPP-based microgrid environment through a multi-process parallel manner, greatly increasing the training speed. Case studies on a real islanded microgrid demonstrate that the proposed method can efficiently and quickly obtain the optimal emergency control strategy while adhering to all hard constraints.

The transition to sustainable power infrastructure necessitates integrating various renewable energy sources efficiently. Our study introduces the deterministic balanced method (DBM) for optimizing hybrid energy systems, with a particular focus on using hydrogen for energy balance. The DBM translates the sizing optimization problem into a deterministic one, significantly reducing the number of iterations compared to state-of-the-art methods. Comparative analysis with HOMER Pro demonstrates a strong alignment of results, with deviations limited to a 5% margin, confirming the precision of our method in sizing determinations. Utilizing solar and wind data, our research includes a case study of Cairo International Airport, applying the DBM to actual energy demands.

This paper proposes a safe soft actor-critic reinforcement learning (RL) algorithm–based controller for networked microgrid restoration. It formulates the post black-start start as a finite-horizon constrained Markov decision process. The RL agent co-optimizes real and reactive power set-points for both grid-forming and grid-following inverters under explicit voltage and frequency constraints, while enforcing proper power sharing via the Mean Active Power Sharing Index (MPSI) and Mean Reactive Power Sharing Index (MQSI). Numerical results obtained on the IEEE 123-bus distribution system show that the proposed method achieves a mean voltage build-up time of 0.01 s without breaching the 5% sharing-violation budget under various load scenarios, considering MPSI and MQSI indices. These findings demonstrate that the proposed method yields fast and safe black-start schedules without resorting to heuristic penalties.

This paper addresses the critical challenge of defending active distribution networks against load-altering attacks via a new Bayesian optimization-enhanced Game theory framework. The attackerdefender interaction is modeled as a finite Stackelberg game, where the attacker strategically manipulates load multipliers to maximize voltage violations and system imbalances, while the defender employs topology control to mitigate these impacts. The framework incorporates both snapshot-based and time-series ramp load-altering attack models, enabling a realistic representation of adversarial behavior that evolves over time or occurs instantaneously. Traditional optimization methods struggle with high-dimensional strategy space and lack adaptability to evolving attack patterns. In our framework, Bayesian optimization operates directly within the game-theoretic formulation to efficiently identify near-optimal attacker and defender strategies under limited computational budgets, outperforming evolutionary baselines in terms of convergence and solution quality. This enables the identification of equilibrium strategies that balance attack effectiveness and defense efficiency. Simulations on the IEEE 123-bus system demonstrate that the framework achieves significant reductions in voltage violations and system imbalances compared to heuristic and learning-based approaches. By optimizing attack and defense strategies, the framework enhances system cyber resilience, offering scalable and adaptive solutions to mitigate modern cyber-physical threats