With increase of practical power system complexity, power system online stability assessment and control is more and more important. Application of the traditional model-driven methods is always limited by contradiction between accuracy and efficiency, while data-driven methods demonstrate strong abilities for the online decision-making support with advancement of various data mining techniques. Instead of direct application of data-driven methods in the power system, this paper first discusses feasible integration approaches for the model-driven and data-driven methods based on the existing achievements, and then, proposes to integrate both methods for the power system online frequency stability assessment and control. The integrated method consists of frequency dynamics prediction and load shedding procedure. In frequency dynamics prediction procedure, integration of system frequency response (SFR) model and the extreme learning machine (ELM)-based learning model is applied, where basic physical causality is kept in the SFR model and ELM is used to fit and correct error of the SFR. The ELM also plays a part in load shedding prediction model construction by digging out mapping relationship from samples. Finally, the proposed prediction and control scheme for the frequency stability is verified by simulations on WSCC 9-bus, New England 39-bus, and NPCC 140-bus system. Results show that the reliability, time efficiency, and accuracy are enhanced with the proposed method.

With the expansion of system scale and data size in power systems, data-driven methods are gradually becoming widely used. However, compared with traditional methods, data-driven methods face more threats in data security and algorithm security. This paper proposes a black box data injection attack method against data-driven strategies in power systems. To obtain a stealthy attack scheme with the greatest impact, an attack vector generator based on deep convolutional neural networks (DCNN) is designed. The generator includes a node selector, a filter, an encoder, and a decoder. The node selector is used to select the most likely successful attack scheme for the filter, the encoder is used to extract sample features, and the decoder is used to generate disturbances based on the features. The proposed generator is trained with an improved generative adversarial network (GAN) and can generate minimal disturbances in real time based on measurement data from the power grid. The effectiveness of the proposed method is demonstrated with an attack experiment on an online transient stability application.

Communication systems serve as the key element in the smart grid as it supports observability and controllability of electric power systems. However, communication failures can lead to an electric power grid operating in an abnormal state causing cascading failures. It is therefore crucial to study adverse effects of communication system failures on power system operation. In this study, a vulnerability assessment framework is proposed to evaluate the role of communication networks on wide‐area power system operation, in which the reliability is evaluated probabilistically considering both latency and communication interruptions; and impact of communication service interruption is evaluated using a multi‐level analytic hierarchy process method. A case study of integrated power and communication systems is used to demonstrate the usefulness of the proposed framework. The proposed method can be used to evaluate the vulnerability of communication systems for electric power grids in both static and dynamic states.

This paper presents a real-time cosimulation platform with hardware-in-the-loop (HIL) for performing cyber-physical power system (CPPS) analyses based on RT-LAB and OPNET software. The platform is capable of simulating a power grid and communication network while also reflecting the impact of actual devices and cyber-attacks on the power system. Furthermore, modelling and implementation methods focused on distributed denial of service (DDOS) and man-in-the-middle (MITM) attacks in a communication network are elaborated. A case study of DDOS and MITM attacks in a typical CPPS is demonstrated based on this platform. The simulation results validate the capability of the platform and also show the importance of considering the separate impacts of the communication system, actual devices and cyber-attack in power system simulations.

Accurate and prompt transient stability prediction is one of the effective ways to reduce the risk of blackout or cascading failures. In an effort to achieve improvements in time efficiency and prediction accuracy, a new transient stability prediction method combining trajectory fitting (TF) and extreme learning machine (ELM) based on two‐stage process, named hybrid method, is proposed here. ELM‐based method is implemented in central station to ensure the time efficiency, while TF‐based method is adopted in local station to guarantee the accuracy. Furthermore, data corruption is taken into consideration to assure the robustness of the proposed algorithm. The hybrid method is validated with the New England 39‐bus test system and the simulation results indicate its effectiveness and reliability.

The rapidly increasing integration of renewable energy sources aggravates the uncertainty and fluctuation in modern power system, which promotes the development of online dynamic security assessment (DSA). Real-time acquisition of high resolution temporal-spatial information on the system states lays the foundation for online DSA, while limited PMU installation and complex dynamic characteristics in real systems impose severe challenges to estimate system states with high quality in real time. This paper proposes a high temporal-spatial resolution state estimation (SE) method, leveraging graph convolutional network (GCN) and dense connectivity structure to estimate states of whole system at PMU reporting rate. Based on proposed SE method, an online DSA framework is developed for transient stability assessment (TSA), which only relies on the hybrid measurements accessible to the control centers in practice. Numerical experiment results in different scenarios demonstrate that the proposed SE method exhibits high SE accuracy and efficiency under different PMU observability. The performance improvement of SE-based TSA approach versus raw-measurements-based TSA approaches is also verified both theoretically and experimentally.