This article proposes a new resilient service restoration approach for a multienergy distribution system (MDS) following unexpected large contingencies. The repair scheduling and MDS restoration are optimally coordinated through a two-stage coordinated framework. To fully exploit the network flexibility during the recovery process, a joint network reconfiguration model is proposed to regulate the switching operations between the power distribution network and the district heating network (DHN). To effectively model the reconfigurable property, the DHN is formulated as a quasi-linear energy flow model for mathematical tractability. Besides, diverse uncertainties from the power, heat, and traffic networks are handled via a two-stage stochastic program. The proposed model is linearized and formulated as a mixed-integer linear programming (MILP) problem considering detailed network operational and temporal-spatial constraints. Furthermore, to accelerate the solution process, a penalty-based Gauss–Seidel decomposition algorithm is developed to decompose and solve the proposed MILP problem efficiently. Finally, comprehensive case studies are done to validate the effectiveness and efficiency of the proposed model and solution method.

Flexible charging and discharging of battery energy storage systems (BESSs) and reactive power compensation of photovoltaic (PV) inverters can be coordinated to support distribution network operation. Given the feature of their rapid response, local droop control is employed to address random fluctuations of renewable outputs and loads. Therefore, this paper proposes a three-stage hierarchical coordination method of BESS-PV scheduling and droop control, aiming to minimize network power loss, voltage deviation, and operating costs. Considering capacity limitation and energy temporal coupling, an optimal state of charge (SoC) interval for one day is determined in a day-ahead stage. Then, theP–Vdroop control functions of BESSs and theQ–Vdroop control functions of PV inverters are optimized hourly, while satisfying the day-ahead SoC interval, to enable efficient real-time local control. Thus, the central day-ahead SoC interval management, the central hourly droop control function scheduling, and the local real-time droop control are hierarchically coordinated. A general model of five-segment droop control functions with reduced binary variables is introduced, and accordingly, a partial segment reduction technique and a penalized linear approximation solution method are developed. The proposed method is tested and compared with other methods. Simulation results verify its high performance in improving the network operational efficiency.

Seaport electrification is widely advocated by the International Maritime Organization (IMO) with the growing worldwide decarbonization trend. The radically increasing demand of the electric vessels has totally reshaped the load profile for onshore microgrid and introduces significant challenges to the seaport operators. However, the modeling of the harbor crafts that provide internal seaport service, especially the coupling between their electricity charging demand and complicated service schedule (such as towage service), has received little attention in the past. Therefore, in this paper, we propose a microgrid operation method to manage the harbor craft fleet with the developed incentives, aiming to reduce the microgrid operating cost while ensuring the fulfillment rates of the internal seaport service. In this method, the charging behaviors of harbor crafts are guided by the adjustable electricity prices, and the broadcast bonus incentivizes the harbor crafts to provide collaborative task support. The proposed model is formulated as a mixed-integer linear program (MILP) problem and solved by an efficient decomposition algorithm. A case study on the Singapore port is conducted with practical data, the dispatch results show that the proposed incentive-based demand response mechanism succeeds in regulating fleet demand and facilitating the utilization rates of charging infrastructures.

This paper proposes a statistical approach for charging demand estimation of an electric harbour craft (e-HC) fleet. The historical operational data of harbour craft is first analyzed and the behaviors of e-HC are estimated accordingly by rule-based methods with detailed technical and practical operational requirements considered. These estimations are then aggregated for estimating the charging demands at a fleet level by Monte-Carlo simulations. With the proposed method, the results of the case study indicate the impacts of charging infrastructure sizing on the operational and economical performances of an e-HC fleet. Sufficient charging capacity improves fleet availability rates and reduces queuing time with higher power demands, while it affects not only operational efficiency but also charging demands as the load of the onshore energy system.

This paper proposes a novel multi-task learning approach based on spatial-temporal recurrent imputation network (SRIN) for power system short-term voltage stability (STVS) assessment with incomplete PMU measurements. The state-of-the-art data imputation methods are based on single and separated learning tasks, which lack optimality for fully exploiting the information in available data. They are also facing several challenges in practical applications, e.g., dependence on complete datasets for training, and performance degradation under continuous data missing scenarios. As a significant advantage, the proposed SRIN method jointly optimizes the objective of missing value imputation and stability prediction through a multi-task recurrent network model. In this way, the integrated model can fully learn from any available data in the incomplete historical database, and the performance of both tasks can benefit from knowledge sharing and transferring across tasks. Moreover, the proposed method has superior advantages in handling both spatial and temporal consecutive missing scenarios, where the imputations are derived by an intelligent combination of history-based and feature-based estimations. Numerical simulation results on two test systems show that, under any PMU missing condition, the proposed method can maintain a competitively high STVS assessment accuracy with a much less imputation error. Note to Practitioners—This paper addresses the challenge of incomplete system observations for power system real-time stability assessment. This problem is not unique to power systems but also extends to other sequential prediction problems facing severe data incompleteness. Existing approaches to solve the missing data problem either relay on complete historical data to train an imputation model, which may not always hold true during practical applications, or impute the missing data by simple statistics, which lacks optimality and adaptivity under diverse missing patterns. This paper proposed a novel, integrated approach to solve this problem by jointly optimizing the two tasks together through a new recurrent network model. In this way, the method can fully learn from seriously undermined datasets. Moreover, this method deals with consecutive missing in time and space, by the design of a trainable weighting component. Numerical simulation results on standard power systems shows that the proposed multi-task model improve the performance of both two tasks and have high adaptivity to different data missing scenarios. In the future research, we will try to address the learning efficiency of this approach for application to larger systems and exploring its adaptability in more extreme scenarios.

For DC microgrids (MGs), real-time adjustment of current sharing ratios and secure voltage restoration are paramount for optimizing load allocation and enhancing dynamic performance. In this paper, a dual-objective distributed model-free adaptive control (MFAC) scheme is designed for the first time to guarantee voltage transient performance and adjustable current sharing. First, an output-constrained nonlinear MG model with ZIP (constant impedance, constant current and constant power) load is established, and subsequently it is converted into an equivalent unconstrained data model using system transformation and dynamic linearization techniques. Second, a new prescribed performance control algorithm with asymmetrical preset boundaries is proposed to restrict voltage transient responses. This algorithm is updated with real-time input and output data at discrete instants, making it independent of line resistance and ZIP load measurements. To enhance the robustness of the control method, an internal observer is designed to actively compensate for the unknown nonlinear dynamics generated by time-varying system parameters. The stability conditions of the transformed systems in the presence of ZIP loads and time-varying line resistance are derived, which can indirectly ensure the prescribed voltage performance of the original system. Finally, the effectiveness of the proposed control method is validated through some simulations and hardware experiments.

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.

This article develops a fully decentralized approximate dynamic programming (FD-ADP) algorithm for stochastic energy management (SEM) of a networked microgrid (NMG) system. First, considering the ac power flow constraints, an alternating direction method of multipliers (ADMM)-based decentralized SEM framework is proposed for NMG coordination. Then, a transactive energy scheme is introduced to further decouple each microgrid (MG) optimization for privacy enhancement and computation reduction. Next, a FD-ADP algorithm is proposed to cope with the real-time uncertainties. The piecewise linear function (PLF) is employed for value function approximation, and a fully decentralized PLF slope update method based on ADMM framework is designed for decentralized property preservation, which trains the value function just through each MG local information and neighboring communication, thus the well-trained decentralized PLF slopes can help achieve the global optimal SEM strategy for NMG coordination under stochastic environments. Finally, case studies demonstrate the effectiveness of the proposed ADMM-based FD-ADP algorithm in terms of decentralized optimization, decentralized training, and global optimality.

Lithium-ion battery health state estimation constitutes an important part of battery management systems, with existing methods either based on mechanistic models or data-driven approaches. This article proposes a physics-informed hybrid multitask learning approach for estimating battery full-life aging states by integrating mechanistic knowledge with data-driven methods at an early lifetime. First, a hybrid aging mode-informed feature is introduced to integrate electrode-level health states with data-driven information. An electrochemical-informed multitask generative model is established to estimate Li$^+$concentration dynamics in both the solid particle and electrolyte. An electrode-level state-constrained training strategy is implemented to guide the model to respect causality. For validation purposes, three battery datasets are utilized to estimate aging states from the electrochemical to the cell level. Compared with traditional mechanistic and data-driven models, the proposed method demonstrates higher accuracy and real-time performance in battery state estimation.

It is generally accepted that the impact of weather variation is gradually increasing in modern distribution networks with the integration of high-proportion photovoltaic (PV) power generation and weather-sensitive loads. This article analyzes power flow using a novel stochastic weather generator (SWG) based on statistical machine learning (SML). The proposed SML model, which incorporates generative adversarial networks (GANs), probability theory, and information theory, enables the generation and evaluation of simulated hourly weather data throughout the year. The GAN model captures various weather variation characteristics, including weather uncertainties, diurnal variations, and seasonal patterns. Compared to shallow learning models, the proposed deep learning model exhibits significant advantages in stochastic weather simulation. The simulated data generated by the proposed model closely resemble real data in terms of time-series regularity, integrity, and stochasticity. The SWG is applied to model PV power generation and weather-sensitive loads. Then, we actively conduct a power flow analysis (PFA) on a real distribution network in Guangdong, China, using simulated data for an entire year. The results provide evidence that the GAN-based SWG surpasses the shallow machine learning approach in terms of accuracy. The proposed model ensures accurate analysis of weather-related power flow and provides valuable insights for the analysis, planning, and design of distribution networks.

Quantum federated learning (QFL) is an emerging interdisciplinary field that merges the principles of quantum computing (QC) and federated learning (FL), with the goal of leveraging quantum technologies to enhance privacy, security, and efficiency in the learning process. Currently, there is no comprehensive survey for this interdisciplinary field. This review offers a thorough, holistic examination of QFL. We aim to provide a comprehensive understanding of the principles, techniques, and emerging applications of QFL. We discuss the current state of research in this rapidly evolving field, identify challenges and opportunities associated with integrating these technologies, and outline future directions and open research questions. We propose a unique taxonomy of QFL techniques, categorized according to their characteristics and the quantum techniques employed. As the field of QFL continues to progress, we can anticipate further breakthroughs and applications across various industries, driving innovation and addressing challenges related to data privacy, security, and resource optimization. This review serves as a first-of-its-kind comprehensive guide for researchers and practitioners interested in understanding and advancing the field of QFL.

Enhanced by machine learning (ML) techniques, data-driven dynamic security assessment (DSA) in smart cyber-physical grids has attracted great research interests in recent years. However, as existing DSA methods generally rely on centralized ML architectures, the scalability, privacy, and cost effectiveness of existing methods are limited. To address these issues, we propose a novel quantum-secured distributed intelligent system for smart cyber-physical DSA based on Federated learning (FL) and quantum key distribution (QKD), namely, quantum-secured federated DSA (QFDSA). QFDSA aggregates the knowledge learned from various local data owners (also known as clients) to predict and evaluate the system stability status in a decentralized fashion. In addition, in order to preserve the privacy of the distributed DSA data, QFDSA adopts the measurement-device-independent QKD, which can further improve the security of local DSA model transmission. Moreover, to accommodate the typical fast system environment and requirement changes, QFDSA alleviates the issues of limited key generation rates by utilizing secret-key pool that guarantee the availability of adequate secret-key materials. Extensive experiments based on the New England 10-machine 39-bus testing system and the synthetic Illinois 49-machine 200-bus testing system demonstrate that the proposed QFDSA method can achieve more advantageous DSA performance while protecting the privacy of local data for real-time DSA applications compared to the benchmarks. Besides, the secret-key generation rate can be improved to adjust its parameters dynamically in real time.

Enhanced by machine learning (ML) techniques, data-driven stability assessment (SA) in smart cyber–physical grids has attracted significant research interest in recent years. However, the current centralized ML architectures have limited scalability, are vulnerable to privacy exposure, and are costly to manage. To resolve these limitations, we propose a novel secure distributed SA method based on federated learning (FL) and differential privacy (DP), namely, Secure Federated SA (SecFedSA). It leverages local system operating data to predict and estimate the system stability status and optimize the power systems in a decentralized fashion. In order to preserve the privacy of the distributed SA operating data, SecFedSA incorporates Gaussian mechanism into DP. Theoretical analysis on the Gaussian mechanism of SecFedSA provides formal DP guarantees. Extensive experiments conducted on the New England 10-machine 39-bus testing system and the synthetic Illinois 49-machine 200-bus testing system demonstrate that the proposed SecFedSA method can achieve advantageous SA performance, while protecting the privacy of the local model information compared to the state of the art.

Voltage stability assessment is essential for maintaining reliable power grid operations. Stability assessment approaches using deep learning address the shortfalls of the traditional time-domain simulation-based approaches caused by increased system complexity. However, deep learning models are shown to be vulnerable to adversarial examples in the field of computer vision. While this vulnerability has been noticed by the power grid cybersecurity research, the domain-specific analysis on the requirements imposed upon effective attack implementation is still lacking. Although these attack requirements are usually reasonable in computer vision tasks, they can be stringent in the context of power grids. In this paper, we conduct a systematic investigation on the attack requirements and credibility of six representative adversarial example attacks based on a voltage stability assessment application for the New England 10-machine 39-bus power system. We show that (1) compromising about half the transmission system buses’ voltage traces is a rule-of-thumb attack requirement; (2) the universal adversarial perturbations regardless of the original clean voltage trajectory possess the same credibility as the widely studied false data injection attacks on power grid state estimation, while the input-specific adversarial perturbations are less credible; (3) the prevailing strong adversarial training thwarts the universal perturbations but fails in defending certain input-specific perturbations. To advance defense to cope with both universal and input-specific adversarial examples, we propose a new approach that simultaneously estimates the predictive uncertainty of any given input of voltage trajectory and thwarts the attacks effectively.

Nonintrusive load monitoring (NILM) aims to analyze the aggregate information of power consumption and recognize the separate operation states of each individual electrical appliance, in which methods of machine learning are frequently used for efficient and effective computation. This article employs multilabel learning models as the main technique to measure the NILM problems with multiple electrical appliances. To precisely extract the information among the aggregate dataset and obtain better effects of data modeling, feature extraction based on interpretation analysis is carried out along of model training. Meanwhile, swapping the order of input labels is also implemented to further characterize the interrelationships and influences among the labels themselves during the modeling process. The results of simulation show that the feature extraction could improve the model performance as it may mitigate the mutual interference among input features and find crucial information to the target electrical appliances. Also, different sort of labels about the operation states of appliances would have impact on the model performance, on both individual and global predictions.

The increasing penetration of renewable energy resources (RESs) has introduced diverse uncertainties in power distribution grids, necessitating the development of advanced grid topology estimation methods applicable in such situation. This article proposes a new method for topology change detection and location tailored for high renewable energy penetrated grids. First, we establish the power grid topology change model using the voltage magnitude measurements, addressing scenarios encompassing line tripping, reconfiguration, and active islanding. We then incorporate the sample covariance model into this framework, elucidating topology changes as low-rank perturbations in the eigenvalue distributions of the voltage matrix. Eventually, topology change detection and location are achieved by monitoring the behavior of the largest eigenvalue within the distribution. Simulation tests are conducted on several distribution power grids, demonstrating the method's efficacy with different levels of RESs penetration. Even in the utmost scenario of 65% RESs integration, our method consistently achieves an impressive topology detection rate of up to 95%. The method also shows high noise tolerance capability and low computational complexity, satisfying the needs for practical application.

The overspeed deloading-based frequency control enables wind turbine generators (WTGs) to provide frequency response. However, the dynamics of WTGs with overspeed deloading-based frequency control under wind speed fluctuations has not been extensively studied. To this end, the dynamics of the Type-3 WTG is studied. In this study, we focus on dynamics of Type-3 WTGs and propose a method for dispatch command tracking through overspeed deloading method under wind speed fluctuations. Then, the counter-effect of the overspeed deloading control on the frequency controller is revealed. To address this issue, a linearization-based method is proposed to improve the frequency response capability of WTGs. Additionally, the pitch control strategy is presented, which can coordinately work with the overspeed deloading-based frequency control. Simulation results show that the proposed method enables WTGs using overspeed deloading to track dispatch command and provide improved frequency regulation support under wind speed fluctuations.

Based on machine learning (ML) technique, the datadriven power system stability assessment (SA) has received significant research interests in recent years. However, even with a high SA accuracy performance, the data-driven SA models may be vulnerable to adversarial examples (caused by some physical noises or adversarial attacks), which are very close to the original input but can result in a wrong SA result. To solve such threat, this paper firstly proposes a universal defense strategy for the MLbased SA models based on randomized smoothing algorithm to resist the adversarial attacks. Secondly, this paper proposes an effectiveness index for the proposed universal strategy to quantify the maximum ability of resistance to adversarial examples. Moreover, this paper provides the tight mathematical proof for the effectiveness index under the hard smoothing, soft smoothing, and binary scenarios. Simulation results verify that the proposed defense strategy can effectively resist the adversarial examples and the proposed effectiveness index can provide formal robustness guarantee for real-time power system SA applications.

Constrained optimization (CO) has made a profound impact in solving many real-world problems. Due to the high computation burden in exact solvers, data-driven CO based on machine learning techniques is recently receiving extensive research interests for its capability to solve CO problems in real time. The existing data-driven CO approaches only serve for optimization problems with rather simple constraints that can be directly incorporated into model training. However, constraints that are computationally infeasible or burdensome to evaluate are commonly experienced in realistic optimization applications, especially in the engineering sector. This paper proposes an athlete–referee dual learning system (ARDLS) for end-to-end CO with large-scale complex constraints, where an athlete model is trained as the main optimizer while a referee model is trained as a probabilistic constraint classifier to guide the athlete training. A risk-based constrained loss function is designed to fine-tune the athlete model for constraint satisfaction. A case study on electric power system emergency control application is conducted to validate the proposed ARDLS, where the testing results demonstrate the excellent capability of ARDLS to improve the likelihood of satisfying large-scale complex constraints in CO.

To provide efficient charging behavior decision-making for urban electric vehicles (EVs), this article proposes a new platform for real-time EV charging navigation (EVCN) based on graph reinforcement learning. Considering the interaction of EVs with charging stations (CSs) and traffic networks, the navigation goal of the “vehicle-station-network” coupled system is to minimize the charging cost and traveling time of EV owners. Specifically, to realize data acquisition and decision-making output, we first characterize the EV charging and traveling behavior as the dynamic interaction process of graph-structured networks. A graph convolutional network is used to extract the environment information required for EVCN, and the generated environment feature is fed into the underlying network of deep reinforcement learning (DRL), which can help the agent better understand massive graph-structured data. Then, the real-time navigation problem is duly formulated as a finite Markov decision process. A sequential scheduling pattern is built according to the sorting of EV charging urgency and solved by a Rainbow-based DRL algorithm. It achieves the sequential recommendation of CSs and planning of traveling routes for multiple EVs. Case studies are conducted within a practical zone in Nanjing, China. Simulation results verify the developed platform and the solving method.

The information exchange among distributed energy resources (DERs) in microgrids (MGs) is through sensing and communication systems, which are prone to expose cyber-attack threats. This article investigates the stability issue of MG systems with distributed secondary control under latency attacks and random denial-of-service (DoS) attacks. Considering these two kinds of attack modes, the corresponding attack consequences including network jamming and time-varying latency in the communication network are simultaneously studied. First, a new metric is defined to quantify the DoS attacks by considering different network jamming choices. Then, the time-domain stability study is conducted considering both attack consequences. Next, a cyber-resilient control strategy is proposed with two control modes: 1) An adaptive-gain resilient controller to sustain the fast stabilization of MG systems under nonuniform time-varying latency attacks, which is proved by the stochastic stability analysis using Lyapunov–Krasovskii functional method. 2) An event-trigger topology reconfiguration controller against excessive latency and damaged cyber connectivity caused by DoS attacks. A switching mechanism for coordinating the above control modes is also designed to guarantee the secondary control functions of MG systems. A modified IEEE 13-bus MG system with five DERs is tested and the effectiveness of the proposed controller under different attack scenarios is verified by OPAL-RT real-time tests.

Due to the stochastic and non-stationary characteristics of wind speed, the wind power generation is highly uncertain and fluctuating, which significantly challenges the operation of the power system and the associated electricity market. In this article, a new spatial-temporal method is proposed for short-term wind power prediction based on image inputs and augmented convolutional network. First, the geographical locations of various wind farms and the relevant wind vectors are processed into a series of multiframe spatial-temporal wind images, which can be handled by the convolutional networks. Then, wind power conversion and prediction models are developed based on those networks, where recurrent paths and attention mechanism are introduced to enhance the model architecture. The testing results have validated the high performance of the proposed method within a forecast horizon of up to seven hours. In particular, even when the terrain information is not available, the implicit wind flow field within the original inputs can still be approximately learned by the proposed convolutional networks.

As a cyber-physical system (CPS), the security of microgrids (MGs) is threatened by unknown faults and cyberattacks. Most existing distributed control methods for MGs are proposed based on the assumption that secondary controllers of distributed generation units operate in normal conditions. However, the faults and attacks of the distributed control system could lead to a significant impact and consequently influence the security and stability of MGs. In this article, a distributed resilient control strategy for multiple energy storage systems (ESSs) in islanded MGs is proposed to deal with these hidden but lethal issues. By introducing an adaptive technique, a distributed resilient control method is proposed for frequency/voltage restoration, fair real power sharing, and state-of-charge balancing in MGs with multiple ESSs in abnormal condition. The stability of the proposed method is rigorously proved by Lyapunov methods. The proposed method is validated on test systems developed in OPAL-RT simulator under various cases.

The optimal deployment of heterogeneous energy storage (HES), mainly consisting of electrical and thermal energy storage, is essential for increasing the holistic energy utilization efficiency of multienergy systems. Consequently, this article proposes a risk-averse method for HES deployment in a residential multienergy microgrid (RMEMG), considering the diverse uncertainties and multienergy demand-side management (DSM). Apart from the HES size and location planning, its optimal investment phase is also determined by maximizing the system equivalent daily profit (EDP) and minimizing the risk. To handle the system uncertainties from renewable energy sources, power demands, outdoor temperature, and residential hot water needs, the multistage adaptive stochastic optimization approach is utilized. Then, through the constraint linearization and stochastic scenario sampling, the original nonlinear deployment model is converted to a mixed-integer linear programming one and tested on an IEEE 33-bus distribution network based RMEMG. The effectiveness of the proposed method is verified by comparing it with the existing practices. The comparison results indicate that the proposed risk-averse deployment method can effectively increase the system EDP and more immune to the uncertainties. Besides, this method can be practically applied for the emerging RMEMGs, such as smart buildings, intelligent homes, etc., which get long-term DSM contracts.

Load shedding (LS) is an effective control strategy against voltage instability in power systems. With increasing uncertainties and complexity in modern power grids, there is a pressing need for faster and more accurate control decisions. In this article, a hierarchical data-driven method is proposed for the online prediction of event-based load shedding (ELS) against fault-induced delayed voltage recovery. The ELS problem is hierarchically modeled as a multi-output classification subproblem for identifying the best shedding location and a regression subproblem to predict the minimum shedding amount. To solve the two subproblems, the weighted kernel extreme learning machine is adopted to construct a direct mapping between the system pre-fault operating conditions and the corresponding control variables. The method is tested on the ELS database, which is analytically generated via a novel adaptive sensitivity-based process on the New England 39-bus system. Compared with other methods, the proposed method is very accurate in prediction with excellent control performance, which maintains superior prediction ability under an imbalanced data distribution.

Internet-of-Things concepts are evolving the power systems to the Energy Internet paradigm. Microgrids (MGs), as the basic element in an Energy Internet, are expected to be controlled in a cooperative and flexible manner. This article proposes a novel distributed control scheme for multiagent systems (MASs) governed MGs in future Energy Internet. The control objectives are frequency/voltage restoration and proportional power sharing. The proposed control scheme considers both intra- and inter-MASs interactions, which offers group plug-and-play capability of distributed generators. The stability and communication delay issues in the control framework are analysed. A multisite implementation framework is presented to explain the agent architecture as well as data exchange in local area networks and the cloud server. Then a cyber hardware-in-the-loop experiment is conducted to validate the proposed control method with multisite implementation. The experimental results prove the effectiveness and application potentials of the proposed approach.

In this article, a novel machine learning based data-driven pricing method is proposed for sharing rooftop photovoltaic (PV) generation and energy storage in an electrically interconnected residential building cluster (RBC). In the studied problem, the energy sharing process is modeled by the leader-follower Stackelberg game where the owner of the rooftop PV system is responsible for pricing self-generated PV energy and operating ES devices. Meanwhile, local electricity consumers in the RBC choose their energy consumption with the given internal electricity prices. To track the stochastic rooftop PV panel outputs, the long short-term memory network based rolling-horizon prediction function is developed to dynamically predict future trends of PV generation. With system information, the predicted information is fed into a Q-learning based decision-making process to find near-optimal pricing strategies. The simulation results verify the effectiveness of the proposed approach in solving energy sharing problems with partial or uncertain information.

This paper describes the PRIMO project which aims at defining, modeling and simulating an operation platform that facilitates market participation of interconnected microgrids (MGs) in the distribution grid market. High penetration of distributed energy resources (DERs) not only increases the complexity of distribution grid operations but can also bring new services and value streams. To this end, interconnected MGs to provide coordination between DERs and distribution grid brings key benefits to improve the social welfare efficiency. The proposed platform comprises physical system modelling tools, and communication interfaces between two hierarchical levels comprising i) distribution grid-to-MG, ii) MG-to-MG energy/flexibility exchange, aiding in the result interpretation and practical realization. Leveraging from the decentralized organization, the operational autonomy and information privacy of each MG is protected.

This paper proposes a cyber-resilience enhancement method to detect and mitigate the denial of service (DoS) attacks in microgrids. It interprets the interactions between cyber and physical systems of microgrids by side-channel detector and dynamic priority scheduling at the first time. It is a cross-layer design scalable to realistic controller and device constrains and compatible with existing designs in both cyber and physical systems. At the same time, it can also guarantee the survival by moving target defense even against the infinite-energy DoS attack at the cost of tolerant performance loss. Numerical simulations verify the effectiveness of the proposed method.

This research paper focuses on coordinated operation scheduling of the condition-based maintenance and energy management of an all-electric ship (AES). The condition-based maintenance comprises of risk assessments, short-duration main-tenance, and condition monitoring of the critical components. The main objective of the proposed strategy is to minimize the fuel, emission, and maintenance cost of the generation and storage units by optimally scheduling the voyage, generation, and risk assessment periods while incorporating the additional information obtained from the condition monitoring equipment. The problem is formulated as a mixed-integer linear programming problem and the non-linear constraints are linearized to improve the computational efficiency. Based on the criticality and risk assessment of the affected units, the operation schedules are continuously updated. The proposed method is validated through a simulation study, and the results demonstrate the applicability and effectiveness of the proposed strategy for the future AES.

This paper proposes a coordinated energy dispatch of the multi-energy ship microgrid with the consideration of the thermal energy requirements of the refrigerated container system (reefer). The proposed combined electricity, cooling, and heating network consists of multiple energy sources (diesel generators, auxiliary electric boiler, auxiliary electric chiller, and gas turbine) and energy storage units (electrical and thermal storage). It aims to minimize the emission cost, the storage operating cost, and startup/shutdown cost of the generation units while maximizing their efficiency. With the help of the piece-wise linearization technique, the coordinated multi-energy dispatch model is formulated as a goal-based multi-objective mixed-integer linear programming problem that can be effectively solved by the commercial solvers. Goal programming (Priori approach) does not require the mapping of the Pareto-front or the solution space and hence saving additional computational power requirements. A case study is carried out to evaluate the effectiveness of the proposed dispatching scheme and the result indicates the improvement in dispatch flexibility by coordinating the electrical power generation with cooling and heating requirements.

This paper designs a novel distributed controller for the islanded DC microgrid. The proposed control method provides a data-driven multi-agent framework to solve the DC bus voltage regulation and current sharing. In order to accurately solve the control action, an online deep reinforcement learning (DRL) algorithm, called deep deterministic policy gradient (DDPG), is employed to secondary controllers in a DC microgrid. Based on the previous knowledge and current system state, DDPG algorithm generates the control action to compensate the voltage reference. In addition, the load reward function for each agent is designed to seek the optimal action of the system. Besides, the proposed control scheme is fully distributed, where each agent only exchange information with neighboring agents. Simulation results of a 4-DG DC microgrid demonstrate the effectiveness and satisfied performance of the proposed multi-agent DDPG-based control strategy.

In this paper, a fully distributed secondary-tertiary coordinated control framework for islanded ac microgrids has been proposed. First, the distributed secondary control for frequency restoration, voltage regulation and power sharing has been proposed. Therefore, the system frequency and bus voltages can be maintained at set points. Each DG can follow the tertiary control signals, while the rest load-generation mismatch is proportionally shared among DGs. Then in the tertiary control, optimal power flow of islanded ac microgrids is formulated as an optimization problem and solved by alternating direction method of multipliers. The entire control framework is achieved in a distributed way with sparse communication networks among each agent. A cyber-physical microgrid platform has been built to validate the proposed controller design in a real-time and hardware-in-the-loop condition. A six-bus three-DG microgrid is developed on the platform and the experimental results validate the effectiveness of the proposed method.

With the deployment of phasor measurement units (PMUs), machine learning based data-driven methods have been applied to online power system stability assessment. This article proposes a novel temporal-adaptive intelligent system (IS) for post-fault short-term voltage stability (STVS) assessment. Unlike existing methods using a single learning algorithm, the proposed IS incorporates multiple randomized learning algorithms in an ensemble form, including random vector functional link networks and extreme learning machine, to obtain a more diversified machine learning outcome. Moreover, under a multi-objective optimization programming framework, the STVS is assessed in an optimized temporal-adaptive way to balance STVS accuracy and speed. The simulation results on New England 39-bus system and Nordic test system verify its superiority over a single learning algorithm and its excellent accuracy and speed without increased computational efficiency. In particular, its real-time assessment speed is 27.5-37.3% faster than the single algorithm based methods. Given such faster assessment speed, the proposed method can enable earlier and more timely stability control (load shedding) for less load shedding amount and stronger effectiveness.

In this article, a novel data-driven method is proposed for open-circuit fault diagnosis of insulated gate bipolar transistor used in three-phase pulsewidth modulation converter. Based on the sampled three-phase current signals, fast Fourier transform and ReliefF algorithm are used to select most correlated features. Then, based on two randomized learning technologies named extreme learning machine and random vector functional link network, a hybrid ensemble learning scheme is proposed for extracting mapping relationship between fault modes and the selected features. Furthermore, in order to achieve an accurate and fast diagnostic performance, a sliding-window classification framework is designed. Finally, parameters in the diagnostic model are optimized by a multiobjective optimization programming model to achieve optimal balance between diagnosis accuracy and speed. At offline testing stage, the overall average diagnostic accuracy can be as high as 99% with the diagnostic time of around one-cycle sampling time. Furthermore, real-time experiments verify its effectiveness and reliability under different operation conditions.

The high penetration of power electronic converter loads in dc microgrid causes system stability issue, or also known as constant power load issue, due to their negative impedance characteristics. The stability concern will be more complicated for a self-disciplined microgrid that allows plug and play of various distributed generations (DGs). This article proposes a robust droop-based controller for decentralized power sharing in a dc microgrid considering large-signal stability. For each DG interface converter subsystem, the interactions with other DG interface converters and loads are estimated by a nonlinear disturbance observer (NDO) utilizing the subsystem's own information to achieve decentralized power sharing and fast voltage regulation. With the uncertainties of circuit parameters modeled as a lumped disturbance term and compensated by an NDO, the proposed controller can significantly enhance the robustness against the uncertainties of circuit parameters. The large-signal stability of the whole interconnected system is proved by the backstepping algorithm and Lyapunov theorem. The efficacy and large-signal stability of the proposed approach are verified by both simulations and experiments.

This article presents a distributed alternating direction method of multipliers (ADMMs) approach for solving the direct current dynamic optimal power flow with carbon emission trading (dc-DOPF-CET) problem. Generally, the ADMM-based distributed approaches disclose boundary buses and branches information among adjacent subsystems. As opposed to these methods, the proposed method (dc-ADMM-P) adopts a novel strategy which uses consensus ADMM to solve the dual of dc-DOPF-CET while only discloses boundary branches information among adjacent subsystems. Moreover, the convergence performance of dc-ADMM-P is improved by reducing the number of dual multipliers and employing an improved update step of the multiplier. DC-ADMM-P is tested on cases ranging from 6 to 1062 buses, with comparison with other distributed/decentralized methods. The simulation results verify the high efficiency of dc-ADMM-P in solving the dc-DOPF problem with complex (nonlinear) factors which can be formulated as convex separable functions. Meanwhile, it also shows the improvement of convergence performance by reducing the number of dual multipliers and employing a new update strategy for the multiplier.

Manufactory load allocation can be used as an effective industrial demand response scheme to reduce operating costs for industrial multienergy microgrids (iMEMGs). In addition, combined cooling, heat, and power (CCHP) plants with auxiliary devices can provide low-cost multiple energies for industrial plants. However, uncertain power generation from renewable energy sources impairs the iMEMG's operation, leading to challenges such as increased operating costs and energy supply deficiency. To conquer these challenges, this paper proposes a multitimescale coordinated adaptive robust operation approach where manufactory load allocation and iMEMG operation are optimally coordinated on different timescales. In the weekly scheduling stage, industrial loads and CCHP units are scheduled for the following week and the hourly iMEMG operation is optimized within the week. Besides, this paper applies an adaptive robust optimization method where the uncertain renewable power generation is fully addressed. The proposed approach is tested on an iMEMG with various industrial manufactories, and it is compared with conventional methods. The simulation results indicate that compared to the conventional ones, the proposed approach can guarantee a robustly optimal operation solution for the iMEMG against any uncertainty realization.

This paper proposes a multitimescale coordinated building energy management system (BEMS) for multienergy buildings integrated with renewable energy sources (RES). It aims to dispatch active building components in two different timescales to counteract uncertain variations in RES generation and load. In the longer timescale (hourly), fuel-cell-based microcombined heat and power and energy storage system (ESS) are dispatched before uncertainty is realized. In the 15 min timescale, ESS is redispatched to supplement the first stage decision, once the uncertainty is realized. The multitimescale coordination is achieved through a two-stage stochastic programming model. The BEMS has been developed as nonlinear receding horizon and nonlinear quadratic programming models in the first and second stages, respectively. In order to minimize carbon footprint of the building, carbon tax has been incorporated in the system operation cost. Furthermore, to prolong battery lifetime in uncertain environment, battery degradation cost has been included. Case studies depict appropriateness of the proposed method in achieving lower carbon emissions while simultaneously improving battery performance in comparison to traditional systems. Extensive simulation results demonstrate robustness and effectiveness of the proposed scheme to account for uncertainties in generation and load.

Fast and stochastic power fluctuations caused by renewable energy sources and flexible loads have significantly deteriorated the frequency performance of modern power systems. Power system frequency control aims to achieve real-time power balance between generations and loads. In practice, it is much more difficult to exactly acquire the values of unbalance power in both transmission and distribution systems, especially when there is a high penetration level of renewable energies. This paper explores a deep learning approach to identify active power fluctuations in real-time, which is based on a long short-term memory recurrent neural network. The developed method provides a more accurate and faster estimation of the value of power fluctuations from the real-time measured frequency signal. The identified power fluctuations can serve as control reference so that the system frequency can be better maintained by automatic generation control, as well as emerging frequency control elements, such as energy storage system. A detailed model of Singapore power system integrated with distributed energy storage systems is used to verify the proposed method and to compare with various classical methods. The simulation results clearly demonstrate the necessity for power fluctuation identification, and the advantages of the proposed method.

This paper proposes an optimal distributed control strategy for the coordination of multiple distributed generators in an islanded microgrid (MG). A finite-time secondary frequency control approach is developed to eliminate the frequency deviation and maintain accurate active power sharing in a finite-time manner. It is demonstrated that the traditional distributed control approach with asymptotical convergence is just a special case of the proposed finite-time control strategy under the specific control parameter settings. Then, a secondary voltage control approach is presented to regulate the average voltage magnitude of all distributed generators to the desired value and achieve accurate reactive power sharing. The implementation of the proposed distributed control strategy only requires information exchange among neighboring local controllers through a sparse communication network. Simulations with an islanded MG testbed built in MATLAB/Simulink are conducted to validate the effectiveness of the proposed distributed control strategy.

As one of the most complex and largest dynamic industrial systems, a modern power grid envisages the wide-area measurement protection and control (WAMPAC) system as the grid sensing backbone to enhance security, reliability, and resiliency. However, based on the massive wide-area measurement data, how to realize real-time short-term voltage stability (STVS) assessment is an essential yet challenging problem. This paper proposes a hierarchical and self-adaptive data-analytics method for real-time STVS assessment covering both the voltage instability and the fault-induced delayed voltage recovery phenomenon. Based on a strategically designed ensemble-based randomized learning model, the STVS assessment is achieved sequentially and self-adaptively. Besides, the assessment accuracy and the earliness are simultaneously optimized through the multiobjective programming. The proposed method has been tested on a benchmark power system, and its exceptional assessment accuracy, speed, and comprehensiveness are demonstrated by comparing with existing methods.

Increasing penetration of plug-in electric vehicles (PEVs) has a substantial impact on the operation of power distribution networks. Given the fast-growing load demands from PEVs and unmatched infrastructure investment in transformer and feeder capacity, the PEV charging is subjected to both spatially and temporally security constraints beyond which the network failure may occur. This paper proposes a game-theory-based distributed charging control method to coordinate large-scale PEVs without compromising the security of the distribution network. Under a noncooperative game framework, a price-driven charging model is designed to minimize the cost of each individual PEV customer while satisfying the network loading constraints. Then, a Newton-type method is developed to find a better Nash equilibrium of the game model at a superlinear convergence rate. Furthermore, an accelerated gradient method is proposed to tackle the subproblem for each user's best response. The update of the user's best response is implemented in a distributed way in order to protect user's privacy. The convergence rate of the proposed algorithms is rigorously proved. The effectiveness and efficiency of the proposed methods are tested on the IEEE 13-bus system.

This paper proposes a new robust load frequency control (LFC) scheme for multiarea power systems based on the second-order sliding mode control and an extended disturbance observer. First, a reduced-order model of the power system LFC is derived. In this model, the load variations and net exchange tie-line power deviations are combined as a lumped disturbance which can be estimated by the extended disturbance observer. Second, a novel sliding surface is designed with the new transformed state variables obtained from the estimated disturbance. The system dynamics can be indicated by sliding surface design using the eigenvalue assignment or the optimal sliding manifold technique. The sliding variable is driven to the sliding surface with a second-order sliding mode algorithm named supertwisting algorithm. The stability of the proposed LFC scheme and the extended disturbance observer is proved using Lyapunov method. The merits of the scheme include faster response speed, stronger robustness against disturbances arising from power system parameter errors, and unmodeled dynamics, and the full consideration of tie-line power flow scheduling variations. Finally, numerical simulations verify the effectiveness of the LFC scheme and reveal its advantages over the state of the arts.

Multienergy microgrids are a promising solution to improve overall energy (electricity, cooling, heating, etc.) efficiency. In this paper, a new optimal energy trading strategy is developed considering the risk from uncertain energy supply and demand in a set of individual multienergy microgrids. According to the historical data about energy supply of each microgrid, an aggregator aims to maximize each microgrid's profit while minimizing the risk of overbidding for renewable energy resources trading based microgrids. A novel two-stage stochastic game model with Cournot Nash pricing mechanism and the conditional value-at-risk criterion is proposed to characterize the payoff function of each microgrid. The sample average approximation (SAA) technique is employed to approximate the stochastic Nash equilibrium of the game model. The existence of the SAA Nash equilibrium is investigated and the corresponding Nash equilibrium seeking algorithm is also realized in a distributed manner. The proposed method is validated by numerical simulations on real-world data collected in Australia, and the results show that the SAA Nash equilibrium based strategy can effectively reduce the risk of not meeting the demand and improve the economic benefits for each microgrid.

In the smart grid and future energy internet environment, a regional energy service provider (RESP) may be able to integrate multiple energy resources such as generator units, demand response, electrical vehicle charging/swapping stations, and carbon emission trading to participate in the market. By imploring a well-known portfolio optimization theory conditional value-at-risk to tackle electricity price uncertainty, this paper formulates the risk-averse day-ahead operational planning for such a RESP as a mixed-integer quadratically constrained programming (MIQCP), named as RA-RESP. A global optimization method, named as multiple perspective-cuts outer approximation method (MPC-OAM) is proposed to solve this model efficiently. A remarkable stronger and tighter mixed integer linear programing master problem is designed to accelerate the convergence of the proposed method. Comprehensive simulation results show that, compared with existing day-ahead planning models, the RA-RESP is a good compromise between profit-based models and cost-based ones. The proposed MPC-OAM can solve complicated RA-RESP problem efficiently, and compared with state-of-the-art solution techniques, the MPC-OAM outperforms in both computing speed and solution quality, especially for scenario which includes more nonlinear factors.

This paper proposes a new robust security-constrained optimal power flow (SCOPF) method to balance the economy. and security requirements under uncertainties associated with renewable generation and load demand. Given the significant growth in microgrid (MG) deployments over the world, this paper explores the potential of using multiple MGs in supporting main grid's security control. Corrective control is employed to relieve postcontingency overflows by effectively coordinating system generators and multiple MGs. An incentive-based mechanism is designed to encourage the MGs to actively cooperate with the main grid for postcontingency recovery, which makes the proposed method to distinguish from the previous models using a traditional centralized control method, such as direct load control. A scenario-decomposition-based approach is then developed to solve the proposed robust SCOPF problem. Numerical simulations on IEEE 14- and IEEE 118-bus systems demonstrate the effectiveness and efficiency of the proposed method.

Dynamic insecurity risk of a power system has been increasingly concerned due to the integration of stochastic renewable power sources (such as wind and solar power) and complicated demand response. In this paper, an intelligent early-warning system to achieve reliable online detection of risky operating conditions is proposed. The proposed intelligent system (IS) consists of an ensemble learning model based on extreme learning machine (ELM) and a decision-making process under a multiobjective programming framework. Taking an ensemble form, the randomness existing in individual ELM training is generalized and reliable classification results can be obtained. The decision making is designed for ELM ensemble whose parameters are optimized to search for the optimal tradeoff between the warning accuracy and the warning earliness of the proposed IS. The compromise solution turns out to significantly speed up the overall computation with an acceptable sacrifice in the accuracy (e.g., from 100% to 99.9%). More importantly, the proposed IS can provide multiple and switchable performances to the operators in order to satisfy different local dynamic security assessment requirements.

Nonintrusive load monitoring (NILM) aims to disaggregate the total power consumption profile measured at the household power inlet into device-level insights. While many studies focus on the modeling methodologies, few of them address the challenge of the computation efficiency which is critical for practical applications. The NILM problem is essentially a nondeterministic polynomial-time hard problem, meaning that obtaining the exact optimal solution is technically intractable. This paper proposes a fast method to address the approximation to the solutions of such problems from an optimization point of view. It is shown that by taking advantage of the constraint programing framework, the computational efficiency of the proposed NILM scheme can be significantly improved while comparable solution accuracy can also be preserved. Simulations conducted on the popular public datasets validate the effectiveness and efficiency of our proposed method.

In the smart grid paradigm, growing integration of large-scale intermittent renewable energies has introduced significant uncertainties to the operations of an electric power system. This makes real-time dynamic security assessment (DSA) a necessity to enable enhanced situational-awareness against the risk of blackouts. Conventional DSA methods are mainly based on the time-domain simulation, which are insufficiently fast and knowledge-poor. In recent years, the intelligent system (IS) strategy has been identified as a promising approach to facilitate real-time DSA. While previous works mainly concentrate on the rotor angle stability, this paper focuses on another yet increasingly important dynamic insecurity phenomenon-the short-term voltage instability, which involves fast and complex load dynamics. The problem is modeled as a classification subproblem for transient voltage collapse and a prediction subproblem for unacceptable dynamic voltage deviation. A hierarchical IS is developed to address the two subproblems sequentially. The IS is based on ensemble learning of random-weights neural networks and is implemented in an offline training, a real-time application, and an online updating pattern. The simulation results on the New England 39-bus system verify its superiority in both learning speed and accuracy over some state-of-the-art learning algorithms.

The architecture of an Advanced Metering Infrastructure with Device Level Load Monitoring (AMI-DLLM) is proposed in this paper. The AMI-DLLM architecture is an upgrade from currently available Advanced Metering Infrastructure (AMI) that is enabled by large-scale smart meter rollouts under National Smart Metering Program (NSMP) across Australia. Potentials of such massive volume of demand side data generated from AMI have not yet been fully explored. The proposed new architecture aims to leverage such affluent demand data from smart meters and enhance both interactivity between utility companies and customers and demand side management. General structure and information flows extended from smart meters in the proposed framework are elaborated in this paper. Discussions of preliminary results and potential applications enhanced by the proposed architecture are also given in detail.

Dynamic security assessment (DSA) is an important issue in modern power system security analysis. This paper proposes a novel pattern discovery (PD)-based fuzzy classification scheme for the DSA. First, the PD algorithm is improved by integrating the proposed centroid deviation analysis technique and the prior knowledge of the training data set. This improvement can enhance the performance when it is applied to extract the patterns of data from a training data set. Secondly, based on the results of the improved PD algorithm, a fuzzy logic-based classification method is developed to predict the security index of a given power system operating point. In addition, the proposed scheme is tested on the IEEE 50-machine system and is compared with other state-of-the-art classification techniques. The comparison demonstrates that the proposed model is more effective in the DSA of a power system.

The increasing penetration of wind power can alter the dynamic security characteristic of a power system. To accommodate rapid and volatile wind power variations, dynamic security assessment (DSA) against foreseeable disturbances is required to be carried out online and provide security monitoring results within sufficiently small time frame. Based on soft computing (SC) technologies, this paper develops an intelligent framework for real-time DSA of power systems with large penetration of wind power. It consists of a DSA engine whose role is to perform real-time DSA of the power system, a wind power and load demand (W&LF) forecasting engine for offline and online predicting wind power generation and electricity load demand, a database generation (DBG) engine for generating instances to train the DSA engine, and a model updating (MU) engine for online updating the DSA engine. Case studies are conducted on two benchmark systems where high DSA efficiency and accuracy are obtained. This framework can be an ideal candidate for advanced security monitoring in the future SmartGrid control centres.