The integration of electric vehicles (EVs) into power systems via vehicle-to-grid (V2G) technology offers new opportunities for bidirectional energy exchange and resource allocation. However, time delays and adversarial attacks in communication networks can undermine the coordination of EV aggregators and power systems. To address this challenge, this paper presents a hybrid counterfactual learning approach for control of EV aggregators in the multi-area power systems V2G and load frequency control (LFC) framework. The proposed hybrid approach integrates counterfactual multi-agent learning, adversarial training, and monotonic neural network (CMA-HMNN). The multi-agent counterfactual learning marginalizes the impact of individual actions on the overall reward, thereby better coordinating controllable resources and reducing variances in adversarial multi-agent training. By enforcing deviation-command monotonicity constraints within the neural network architecture, the proposed approach can satisfy Lyapunov stability conditions and improve the stability of power systems with integrated EVs. Adversarial training based on the fast gradient sign method (FGSM) is applied to enhance the robustness of the networks against perturbations. Even under concurrent time-varying communication delays and malicious threats from communication networks, the method effectively coordinates multiple generation resources and EV aggregators. Compared with four DRL-based control methods, the superiority of the proposed method is verified on the three-area power system and IEEE 39-bus power system with wide EVs integrations.

Service restoration implemented by multiple distributed energy resources (DERs) is a resilience-enhancing paradigm for modern distribution systems. To address the challenges of complex system modeling and the problem of cooperative control over heterogeneous multiple agents, this article proposes a graph reinforcement learning (G-RL) method based on heterogeneous multiagent systems (MASs). The method leverages graph-structured data to enhance the representation of distribution system states and employs graph attention networks (GATs) to deeply explore the power flow features and spatial characteristics of nodes in the restoration process. Additionally, a multihead self-attention (MHSA) is incorporated to strengthen collaboration among heterogeneous agents, enabling them to focus on relevant information from multiple perspectives during training. Finally, a joint simulation test platform is developed using Python and OpenDSS, and case studies on a 123-bus distribution system are conducted. Experimental results demonstrate that the proposed approach achieves efficient and autonomous service restoration by enhancing spatial feature extraction and improving collaborative decision-making among agents.

The large-scale integration of distributed energy resources into active distribution networks may significantly intensify voltage fluctuations and increase network losses. Traditional model-based reactive power optimization approaches depend on existence of accurate system models. On the other hand, conventional reinforcement learning methods largely ignore the spatial characteristics of the active distribution networks during training, allowing agents to have an inadequate perception of the system state. To address these challenges, this paper proposes a multi-agent deep reinforcement learning approach that integrates graph learning with reinforcement learning for the learning of reactive power optimization strategies in active distribution networks. Specifically, the active distribution network is divided into multiple regions, with each region being controlled by an agent. The agents collaborate to achieve the global reactive power optimization goal. The perception capability of the agents is enhanced by adopting graph attention networks during the feature extraction phase. In the training phase, a primal-dual method is employed to manage constraints effectively. During the execution phase, each agent controls the photovoltaic inverters, electric springs, and capacitor banks based on the strategies developed in the training phase. The performance of the proposed approach is validated by a series of experiments on the IEEE-33 system, along with comparisons versus some existing data-driven deep reinforcement learning methods.

With the development of intelligent automation technology and advancement of modernization, the degree of interconnection between power systems is increasing. With the main purpose of involving hybrid energy storage systems (HESS) in optimizing system frequency, this work proposes a learning-based tube model predictive control (MPC) for the multi-area interconnected power systems with wind power and HESS. The suggested method has strong adaptability due to the introduction of a new robust constraint handled by a learning mechanism. By identifying the uncertainty set of coupling strength of online data in the learning stage, the optimal MPC problem is calculated in the adaptive stage, which effectively reduces the adverse effects of disturbances and noises in multi-area interconnected power systems. Moreover, an input to state stability criterion is provided to ensure the robust stability of the system with uncertain disturbances and noises. With simulations on a four-area interconnected power system with wind power and HESS, the effectiveness of proposed method is discussed on an improved IEEE 39-bus system.

Lithium-ion battery charging involves many factors such as electricity, heat, aging, etc. Shortening the charging time of lithium-ion batteries while limiting aging and temperature rise is an urgent issue that needs to be addressed. In order to solve this problem, the pseudo-two-dimensions (P2D) model is used to describe the electrical behavior of batteries. Then, a parameter identification method is proposed using a deep deterministic policy gradient for P2D model. Besides, a multi-objective charging model considering time, health loss, and temperature rise is constructed by combining P2D, equivalent heat, and aging mechanism models. A multi- objective charging optimization strategy is proposed using Deep Q-Network algorithm. Under the condition of strong coupling of multiple parameters in the P2D model, the identification accuracy is improved by more than 20% compared to traditional optimization algorithms. Under fast charging conditions, the charging time decreases by 19%, and the health loss only increased by 2.5%. Finally, the impact of multi-stage current variance on battery health loss is discussed.

A memory output sliding mode load frequency control (MOSMLFC) strategy is proposed for multi-area interconnected power systems under historical-frequency-triggered denial-of-service (DoS) attacks. Due to the use of the open network, the multi-area power system is prone to cyber-attacks. Different types of attack models have been built to describe the actual attack behavior, so that effective strategies can be quickly formulated in the event of an attack. Therefore, a historical-frequency-triggered DoS attacks model is presented from the perspective of attackers, with the aim of destroying the stable state of the multi-area power system. It is assumed that attackers determine the timing of DoS attacks by monitoring the operational status of multi-area power systems and designing the triggering condition with historical frequency. A MOSMLFC strategy is investigated to ensure the security performance of multi-area power systems under historical-frequency-triggered DoS attacks, which applies the memory output information of the power system to realize the controller design. The security condition of multi-area power systems under historical-frequency-triggered DoS attacks is obtained by Lyapunov’s theorem and linear matrix inequality (LMI). Numerical examples are tested over the IEEE 10-generator 39-bus system and the results prove the usefulness and superiority of the proposed method. Note to Practitioners—Load frequency control is widely applied in multi-area power systems to achieve a balance between the load demand and generation. Frequent cyber-attacks are a threat to the normal operation of the power system. It is therefore necessary to develop appropriate strategies to defend against cyber-attacks. So far, there have been many different forms of cyber-attacks. This has prompted defenders to build different types of attack models to describe the actual attack behavior in order to preemptively formulate appropriate defensive strategies. Smart attacker may notice that certain characteristics of the target system are important, such as the power system frequency. This motivates us to propose a historical frequency-triggered DoS attack model that contributes to a deep understanding of the impact of cyber-attacks on the power system. We propose a unique sliding mode control approach to ensure the stable performance of power system state and output simultaneously.

For voltage sag position tracing (VSPT) through deep learning methods, model performance deteriorates rapidly under changed scenes. Moreover, time and effort are wasted in retraining numerous models for all different scenes. Therefore, a self-adaptive VSPT method which can response to changed scene is urgently needed. In this article, a deep transfer learning for self-adaptive VSPT under changed scenes is proposed. For accurate VSPT under original scene, a deep learning method via temporal iTransformer is presented, which can enhance local feature extraction capability while retaining the iTransformer’s global perspective. For self-adaptive VSPT under changed scenes, a deep transfer learning based on feature-decoupling is further presented. Here, domain invariant features are calculated via feature-decoupling module, and the difference between source domain features and target domain features is adaptively minimized via feature transference. We test the proposed method via simulation and experimental platform, verifying that the proposed deep transfer learning has satisfactory domain adaptability for self-adaptive VSPT under changed scenes.

A reliable adaptive-memory-derivative (AMD) event-triggered quantized sliding mode load frequency control (QSMLFC) method is proposed for the multiarea interconnected wind power system under frequency-based deception attacks. An AMD event-trigger scheme is proposed to promote the wind power system operation while saving the network resources, and the reliable AMD event-triggered QSMLFC method aims to reduce the frequency deviations of the interconnected wind power systems. A frequency-based deception attack model is developed for analyzing the security issues in network communications for wind power systems. The hysteresis quantizer is used to lower the communication rate. To validate the correctness of the control method, a sufficient reliability criterion is derived to prove the applicability of the AMD event-triggered QSMLFC. Three numerical examples and an IEEE 39-bus system simulation are presented to demonstrate that the reliable AMD event-triggered QSMLFC method can provide satisfactory stability performance for the wind power system under frequency-based deception attacks.

With the advancement of artificial intelligence, deep reinforcement learning is emerging as an effective solution for distribution system service restoration. However, traditional deep reinforcement learning approaches are typically tailored for training agents in specific scenarios, limiting their ability to adapt rapidly to new environments. Furthermore, the spatial characteristics of the distribution network are largely ignored during the training, constraining the state perception capabilities of agents. To address these issues, this paper proposes a meta-based graph reinforcement learning approach that combines graph learning, meta-learning, and reinforcement learning for the learning of service restoration strategies in distribution network. The agent trained by such an approach possesses the feature perception capability of graph learning, allowing it to acquire deeper service restoration strategies from latent graph features. Moreover, the agent also has the fast adaptation ability of meta-learning, enabling it to quickly adapt to new restoration scenarios. Experimental results demonstrate that the proposed approach outperforms existing results of both specialized and generalized strategies.

This paper studies the dynamic state estimation problem for cyber-physical distribution systems (CPDSs) with false data injection attacks (FDIAs) and delayed measurements. In view of the characteristics of multiple measurement types, the equivalent current measurement transformation technique is adopted to make the measurement equation be expressed in the form of linear measurement model. Based on the mixed measurements of phasor measurement units (PMUs) and distribution remote terminal units (DRTUs), a novel model is constructed using Bernoulli distributed random variables to describe the delay phenomena. Further, in order to improve the transmission efficiency of the measurement data, an mechanism is introduced in the network transmission process to minimise the amount of data transmission in the network while ensuring the performance of system state estimation. A measurement model based on the event-triggered mechanism is developed, and an event-triggered square root cubature Kalman filter (ET-SRCKF) algorithm incorporating delayed measurements is designed to implement the state estimation of CPDSs, which can obtain the optimal estimation of the states under delayed measurements. Finally, simulated examples are conducted on the IEEE 33-bus test system, and the effectiveness of the proposed method is illustrated by numerical simulations.

This paper presents a novel non-fragile output sliding mode load frequency control (OSMLFC) strategy designed for multi-area interconnected power systems that incorporate electric vehicles (EVs), particularly in the presence of frequency-triggered denial-of-service (DoS) attacks. We delve into the realm of network communication security concerning load frequency control (LFC) power systems combined with EVs, investigating a real-time frequency-triggered DoS attack by combining real-time frequency dynamics with event-triggering mechanisms. A non-fragile output sliding mode control (SMC) method is proposed, strategically devised to balance the load and frequency aspects of the power systems. Then, a sufficient stability criterion is derived to ensure the non-fragile$H_\infty$stability of the power system integrated with EVs, even when subjected to the perturbations caused by real-time frequency-triggered DoS attacks. The efficacy of our proposed approach and the characteristics of the real-time frequency-triggered DoS attacks are validated through extensive simulations.

This article investigates the resiliently distributed fixed-time control of frequency recovery and power allocation in a multi-terminal high voltage direct current (MTDC) system against denial-of-service (DoS) attacks. An MTDC system typically consists of several AC areas, on which the DoS attacks may cause communication faults by blocking communication channels, preventing certain AC areas from sending message and damaging related facilities. A novel distributed security control scheme is proposed in this paper, which introduces attack detection method and communication repair mechanism to restore the paralyzed topology caused by DoS attacks. By extension, a resiliently distributed fixed-time control is presented under this frame. The proposed control scheme can not only realize frequency restoration but also accomplish active power sharing under DoS attacks. Furthermore, different from existing control strategies, the advanced scheme can guarantee the convergence time without considering the initial value, which helps improve the robustness and stability of the MTDC system. The resilient stability of the proposed scheme is proved by Lyapunov-Krasovskii stability theory. Finally, case studies on an MTDC system are conducted to demonstrate the effectiveness and validity of the proposed controller. Note to Practitioners—MTDC system is a large-scale power system connecting various AC grids. It has the characteristics of distributed and high intelligence, which is prone to be attacked by an adversary. As an index to measure the safe and stable operation of MTDC system, frequency is the focus of this paper. We propose a novel topology recovery mechanism for MTDC systems under DoS attack and design a resilient fixed-time secondary frequency controller based on the idea of multi-agent. The experimental results show that under DOS attack, the proposed topology recovery mechanism and controller can recover the frequency to the rated value in a fixed time and realize the proportional distribution of active power. In practical application, engineers can learn from the controller to resist DoS attack and realize the stable operation of large-scale distributed power system.

The large-scale grouping of the battery system leads to the inconsistency of the battery pack. Aiming at tacking this issue, an inconsistency evaluation method is deployed for the battery pack based on an improved Gaussian mixture model (GMM) and feature fusion approach. Specifically, the proposed adaptive forgetting factor recursive least squares (AFFRLS) algorithm allows the open-circuit voltage and other parameters to be jointly identified without the open circuit voltage-state of charge (OCV-SOC) test. An online capacity estimation approach with the extended Kalman particle filter (EPF) is put forward for capacity estimation. Further, an improved GMM is proposed to visualize battery pack inconsistency, using the K-means++ algorithm to initialize category centers. The standard deviation coefficient approach quantifies the inconsistency. Finally, the real-life vehicle data are performed to validate the effectiveness of the proposed method. The experimental results show that the proposed method can evaluate the battery parameters accurately. With the increase in service time, the inconsistency of the battery pack is gradually deteriorating.

Abstract This paper proposes a state estimation approach ‘robust strong tracking unscented Kalman filter with unknown inputs’ that can be applied to non‐linear systems with unknown inputs. Specifically, the non‐linear state and measurement equations are linearised by statistical linearisation. Then, the estimation equation of the unknown input is derived based on the weighted least squares method. The multiple suboptimal fading factor is introduced into a priori error covariance matrix to improve the tracking ability for the inaccuracy of the system model and the abrupt change of state variables caused by unknown inputs. Finally, based on the unbiased minimum variance estimation, the unbiased state estimation and the error covariance matrix are derived. Singular value decomposition is performed on the error covariance matrix to improve the stability of the algorithm. Simulated results validate the effectiveness of the proposed method.

Considering various cyberattacks aiming at the Internet of Vehicles (IoV), secure pose estimation has become an essential problem for ground vehicles. This paper proposes a pose estimation approach for ground vehicles under randomly occurring deception attacks. By modeling attacks as signals added to measurements with a certain probability, the attack model has been presented and incorporated into the existing process and measurement equations of ground vehicle pose estimation based on multisensor fusion. An unscented Kalman filter-based secure pose estimator is then proposed to generate a stable estimate of the vehicle pose states; i.e., an upper bound for the estimation error covariance is guaranteed. Finally, the simulation and experiments are conducted on a simple but effective single-input-single-output dynamic system and the ground vehicle model to show the effectiveness of UKF-based secure pose estimation. Particularly, the proposed scheme outperforms the conventional Kalman filter, not only by resulting in more accurate estimation but also by providing a theoretically proved upper bound of error covariance matrices that could be used as an indication of the estimator’s status.

High-voltage direct current (HVDC) is increasingly utilized for long-distance electric power transmission, mainly due to its low resistive losses. In this paper, a decentralized control strategy is proposed to address the secondary frequency restoration and real power sharing problem for multi-terminal direct current (MTDC) transmission systems. We establish a sufficient stability condition to guarantee that the designed decentralized leaky integral controller can restore the frequency to its nominal value. Furthermore, the proposed controller can adjust the real power sharing ratio according to different working conditions. An MTDC system consisting of 4 AC systems is built in MATLAB Simulink environment. Numerical simulations are conducted to validate the effectiveness of proposed decentralized control approach.

We consider the quasi-synchronization problem of a continuous time generalized Markovian switching heterogeneous network with time-varying connectivity, using pinned nodes that are event-triggered to reduce the frequency of controller updates and internode communications. We propose a pinning strategy algorithm to determine how many and which nodes should be pinned in the network. Based on the assumption that a network has limited control efficiency, we derive a criterion for stability, which relates the pinning feedback gains, the coupling strength, and the inner coupling matrix. By utilizing the stochastic Lyapunov stability analysis, we obtain sufficient conditions for exponential quasi-synchronization under our stochastic event-triggering mechanism, and a bound for the quasi-synchronization error. Numerical simulations are conducted to verify the effectiveness of the proposed control strategy.

This paper considers the problem of asynchronous H∞control for networked Markovian jump systems subject to probabilistic packet dropouts and communication delays in the measurement channel. A new dynamic output-feedback-based asynchronous controller is proposed wherein the dynamic output-feedback controller modes need not synchronize with the system modes. By utilizing results from stochastic Lyapunov-Krasovskii stability theory, sufficient conditions in terms of matrix inequalities are derived such that the closed-loop networked Markovian jump system is stochastically stable and achieves the prescribed H∞performance. Using the Schur complement technique and under the assumption that the input matrix is full rank, the sufficient condition is reduced to a linear matrix inequality and the dynamic output-feedback-based asynchronous controller is synthesized. A detailed numerical example with simulation results are presented to evaluate the proposed controller design scheme.

Due to the existence of various anomalies such as non-Gaussian process and measurement noises, gross measurement errors, and sudden changes of system status, the robust forecasting-aided state estimation is pivotal for power system stability. This paper develops a novel unscented Kalman filter (UKF) with the generalized correntropy loss (GCL) (termed as GCL-UKF) to estimate power system state with forecasting aid. The GCL is used to replace the mean square error loss in the original UKF framework. The advantage of such an approach is that it combines the strength of the GCL developed in robust information theoretic learning for addressing the non-Gaussian interference and the strength of the UKF in handling strong model nonlinearities. In addition, we take into account the nontrivial influences of the bad data for the innovation vector. An enhanced GCL-UKF method is established by introducing an exponential function of the innovation vector to adjust a covariance matrix so as to improve the GCL-UKF-based state estimation accuracy under the change of gain matrix caused by bad factors. Numerical simulation results carried out on IEEE 14-bus, 30-bus, and 57-bus test systems validate the efficacy of the proposed methods for state estimation under various types of measurement.