Wide-area Blackouts Research Articles

A multi-agent reinforcement learning method for distribution system restoration considering dynamic network reconfiguration

Extreme weather, chain failures, and other events have increased the probability of wide-area blackouts, which highlights the importance of rapidly and efficiently restoring the affected loads. This paper proposes a multi-agent reinforcement learning method for distribution system restoration. Firstly, considering that the topology of the distribution system may change during network reconfiguration, a dynamic agent network (DAN) architecture is designed to address the challenge of input dimensions changing in neural network. Two encoders are created to capture observations of the environment and other agents respectively, and an attention mechanism is used to aggregate an arbitrary-sized neighboring agent feature set. Then, considering the operation constraints of the DSR, an action mask mechanism is implemented to filter out invalid actions, ensuring the security of the strategy. Finally, an IEEE 123-node test system is used for validation, and the experimental results showed that the proposed algorithm can effectively assist agents in accomplishing collaborative DSR tasks.

A Method for Variance-Based Sensitivity Analysis of Cascading Failures

Cascading failures of relay operations in power systems are inherently linked with the propagation of wide-area power system blackouts. In this paper, we consider a power system cascading failure as an indicator matrix encoding: what power system relays operated within a cascading failure inherently capturing the component and the sequence of tripping events. We propose that this matrix may then be used with extended forms of variance-based sensitivity estimators to quantitatively rank how sensitive observed power system cascading failures are to power system variables, considering overall system cascading failures as well as cascading failures grouped by network area and relay types. We demonstrate our proposed method by investigating the sensitivity of cascading failures to relay parameters, system conditions, and fault location using a version of the IEEE 39 bus model modified to include protection relays, wind farms, and tap-changing transformers. Input power system variables included: system operational scenario, disturbance location, relay parameters or thresholds. The Case Studies' results confirm the method's utility by successfully generating relative rankings of input variables' importance with respect to cascading failure propagation. The results also show cascading failures' sensitivity to input variables to be high due to non-linear relationships between input variables and cascading failures.

재생에너지 기반 Hybrid 전력망의 주파수 추정

The frequency is one of the very important factors in power grid. It is essential that the frequency of a power grid should be maintained close to its nominal frequency. Recently, under the leadership of the government, interest in the expansion and distribution of renewable energy sources (RES) is increasing. Therefore, technology for stable operation in preparation for grid connection variability of RES is required. There is growing interest in the rapid measurement and accurate estimation of grid frequency for renewable energy monitoring and stability, using GPS based time-synchronous PMU (Phasor Measurement Unit) for wide area blackout prevention. In this paper, FRDWT (Fast Recursive Discrete Wavelet Transform), GCDFT (Gain Compensator Discrete Fourier Transform) and PAD (Phase Angle Difference) are selected as optimal frequency measurement methods for frequency estimation of hybrid power grid based on renewable energy. This study is to evaluate the performance of the proposed technique using domestic 345kV power grid model data by EMTP-RV S/W.

Grid-forming control strategies for black start by offshore wind power plants

Abstract. Large-scale integration of renewable energy sources with power-electronic converters is pushing the power system closer to its dynamic stability limit. This has increased the risk of wide-area blackouts. Thus, the changing generation profile in the power system necessitates the use of alternate sources of energy such as wind power plants, to provide black-start services in the future. However, this requires grid-forming and not the traditionally prevalent grid-following wind turbines. This paper introduces the general working principle of grid-forming control and examines four of such control schemes. To compare their performance, a simulation study has been carried out for the different stages of energization of onshore load by a high-voltage direct-current (HVDC)-connected wind power plant. Their transient behaviour during transformer inrush, converter pre-charging and de-blocking, and onshore block-load pickup has been compared and analysed qualitatively to highlight the advantages and disadvantages of each control strategy.

Probabilistic Study of Undervoltage Load Shedding Scheme to Mitigate the Impact of Protection System Hidden Failures

The investigations on recent blackouts have disclosed that one of the main reasons for cascaded tripping of protection systems is hidden failures (HFs). This involves an in-depth study of the different modes of HF in protection systems to assess their impact on cascading events in a power grid. This paper portrays a detailed study on the modeling of different HF modes and proposes a supervised undervoltage load-shedding scheme to further reduce the likelihood of cascading events caused by HFs. The proposed methodology uses two different probabilistic models of HFs to assess HFs’ impact on cascading events. The proposed methodology is validated through numerous case studies on the IEEE 57-bus network. It is also observed that inclusion of self-monitoring and control features in digital relays significantly reduces the risk of wide-area blackouts. The methodology can be effectively adopted for planning, inspection, and repair of the protection system and the maintenance of it where investment guarantees its reliability.

Analysis of Possibility of Replacing Self-starting Generator with BESS

Recently, Many papers have been published on other's utility of the existing BESS (Battery Energy Storage System) equipment that stabilizing the output of the renewable energy and adjusting the frequency of the power system. In 2013, the EPRI has reported a study of the primary power plant starting using BESS for the replacement of two black start generators of American Electric Power (AEP) that will be torn down. As the social interest in the power outage has increased since wide-area blackout in South Korea in 2011 it is time to research not only a reliability of the power restoration plan but also reducing power restoration time. In case of wide area black-out, the primary transmission line should be energized at first after opening all circuit breaker in power system. Since these transmission lines are long-distance and no-load, various problems may befall in the process of energizing. In this paper, the restoration process using BESS for primary transmission line was analyzed. PSCAD / EMTDC was used for simulation.

A Three-Stage Procedure for Controlled Islanding to Prevent Wide-Area Blackouts

Controlled islanding has been proposed as a last resort action to stop blackouts from happening when all standard methods have failed. Successful controlled islanding has to deal with three important issues: when, and where to island, and the evaluation of the dynamic stability in each island after islanding. This paper provides a framework for preventing wide-area blackouts using wide area measurement systems (WAMS), which consists of three stages to execute a successful islanding strategy. Normally, power system collapses and blackouts occur shortly after a cascading outage stage. Using such circumstances, an adapted single machine equivalent (SIME) method was used online to determine transient stability before blackout was imminent, and was then employed to determine when to island based on transient instability. In addition, SIME was adopted to assess the dynamic stability in each island after islanding, and to confirm that the chosen candidate island cutsets were stable before controlled islanding was undertaken. To decide where to island, all possible islanding cutsets were provided using the power flow (PF) tracing method. SIME helped to find the best candidate islanding cutset with the minimal PF imbalance, which is also a transiently stable islanding strategy. In case no possible island cutset existed, corresponding corrective actions such as load shedding and critical generator tripping, were performed in each formed island. Finally, an IEEE 39-bus power system with 10 units was employed to test this framework for a three-stage controlled islanding strategy to prevent imminent blackouts.

Power grid transient stabilization using Koopman model predictive control

Abstract This work addresses the problem of transient stabilization of a power grid, following a destabilizing disturbance. The model considered is the cascade interconnection of seven New England test models with the disturbance (e.g., a powerline failure) occurring in the first grid and propagating forward, emulating a wide-area blackout. We consider a data-driven control framework based on the Koopman operator theory, where a linear predictor, evolving on a higher dimensional (embedded) state-space, is built from observed data and subsequently used within a model predictive control (MPC) framework, allowing for the use of efficient computational tools of linear MPC to control this highly nonlinear dynamical system.

A New Wide Area Intelligent Multi-agent Islanding Detection Method for Implementation in Designed WAMPAC Structure

In this paper, a new multi-agent distributed neural network based islanding detection algorithm is presented. The proposed algorithm is placed in the designed wide area monitoring, control, and protection system (WAMPAC) as one of its applications. Therefore, firstly, the proposed WAMPAC structure and its main parts including WAMS, WACS and WAPS are highlighted that are designed to have a situational awareness to perform both of control and protection actions. To have fast and accurate responses at different contingencies such as WAMS equipment failures or data transmission latency, the multi-agent concept, and neural network algorithm are used at some of the WAMPAC control and protection applications. To show the effectiveness of the designed algorithms, the islanding detection application and its algorithms are described in detail. The proposed method is applied to the IEEE-118 bus power system and its performance is demonstrated at various disturbances. It is shown that the islanding necessity is predicted correctly that helps to perform corrective actions such as controlled islanding execution to avoid wide area blackout. In addition, the algorithm response is increased significantly that contributes to overcome the possible drawbacks of WAMS in large interconnected power systems.

Adaptive Protection and Control in the Power System for Wide-Area Blackout Prevention

This paper presents a new approach in adaptive out-of-step (OOS) protection settings in power systems in real time. The proposed method uses extended equal area criterion (EEAC) to determine the critical clearing time and critical clearing angle of the system, which are vital information for the OOS protection setting calculation. The dynamic model parameters and the coherency groups of the system for EEAC analysis are determined in real time to ensure that the newly calculated settings suit with the prevalent system operating condition. The effectiveness of the method is demonstrated in simulated data from a 16-machine 68-bus system model.

A Bayesian Algorithm to Enhance the Resilience of WAMS Applications Against Cyber Attacks

Monitoring critical infrastructures is highly dependent on the accuracy of the installed sensors and the robustness of estimation algorithms. Data-injection attacks can degrade the operational reliability and security of any cyber-physical infrastructures. An attacker can compromise the integrity of the monitoring algorithms by hijacking a subset of sensor measurements and sending manipulated readings. Such approach can result in wide-area blackouts in power grids. This paper considers several cases of severe data-injections with high probabilities of information loss. To achieve an accurate supervision, a Bayesian-based approximated filter (BAF) has been proposed at each monitoring node using a distributed architecture. To maintain a reduced communication overhead and time complexity, upper and lower bound methods have been developed. The performance of the proposed technique has been demonstrated in a mature synchrophasor application known as the oscillation detection. Two test cases have been generated to examine the immunity of the proposed estimation scheme in New Zealand and Oman power grids. The tests were conducted in the presence of harsh data-injection attacks and multiple system disturbances. Results show the proposed BAF method can accurately extract the oscillatory parameters from the contaminated measurements.

Optimal design of convertible static compensator supplementary damping controller to avoid wide area uncontrolled islanding

This study presents a new design of a wide area damping controller for the convertible static compensator (CSC) to limit some major blackouts due to asynchronous oscillations. The adaptive neuro-fuzzy inference system (ANFIS) approach is adopted in the dual heuristic programming method to have an optimal response at different disturbances. The proposed ANFIS parameters are optimised by using the multi-swarm particle swarm optimisation method. The CSC dynamic performance in both of unified power flow controller and interline power flow controller configurations is investigated to increase the damping of inter-area oscillations and to prevent the uncontrolled network islanding. The Iranian power grid and New England power system are selected to install CSC and to apply the designed damping controller. It is shown that the considerable number of wide area blackouts could be avoided by CSC application equipped with the proposed damping controller. In other scenarios, which the controlled islanding is an unavoidable task, the boundaries of islands are determined adaptively considering the changes of generators coherency and the loadings of lines. Simulation results verify that the load shedding amounts are decreased and the stability margins of islands are increased. The uncontrolled islanding condition is also delayed significantly.

A knowledge based decision support algorithm for power transmission system vulnerability impact reduction

Abstract One of the main reasons for wide area blackouts is cascading failures due to critical contingencies. Since, several factors such as faults, misoperations, environmental effects and sabotage issues are involved; the analysis of such contingencies is a challenging process. Most available methods in literature deal mainly with a certain aspect of the problem. In this study, a more comprehensive methodology using operational and non-operational indices for power system vulnerability analysis is presented. The individual indices are defined for power system’s operational performance, terrorist attack and adverse weather conditions where a fuzzy inference system is used to obtain a single Total Vulnerability Index for each transmission system line. Additionally, a knowledge based decision support algorithm is proposed for the use of TSO’s defense activities in order to determine the weak points and counter measures like load shedding. The modeling approach is tested on IEEE reliability test network with imposed external constraints.

New intelligent controlled islanding scheme in large interconnected power systems

This study presents a new intelligent controlled islanding scheme based on wide area measurement systems data to avoid the wide area blackout. Three offline, online and real-time parts are applied to solve three problems including where and when to implement islanding and what to do after separation. New security-based criteria are used to determine the initial stable coherent groups. The boundaries of islands are obtained adaptively considering different operating points by using the weighted time varying graph structure of the network. To reach more stable islands, reactive power is considered by using a self-tuned online fuzzy factor in graph weights. The number of necessary islands with their locations is determined in online part by monitoring the dominant inter-area oscillations between the initial groups (IGs). Then, the network is split into islands with the objective of minimum power flow disruption. To detect the unavoidable islanding cases correctly, a new parallel adaptive neuro-fuzzy inference system (ANFIS) structure is designed. In a parallel structure, for each of two adjacent IGs a distinct ANFIS is also applied to consider variable stability margins between groups. Simulation results confirm that the blackout can be avoided in a large power grid by using the proposed method.

Decision support for restoration of interconnected power systems using tie lines

Existing system restoration strategies are primarily based on operators’ experience with no specific decision support tools. Restoration of interconnected grids in Europe is decentralised; restoration planning is limited to within the area of control for each national transmission system operator (TSO). Although coordination of restorative actions between TSOs is encouraged, few agreements exist. This study proposes a decision support methodology for system operators and restoration planners for restoration of interconnected grids. Solutions to optimal and coordinated utilisation of tie lines (TLs) are proposed. The proposed tool facilitates collaboration between TSOs affected by wide area blackouts. The decision support tool allows to adaptively develop restoration strategies based on system conditions, available resources and operating constraints. Optimisation algorithms are proposed to determine the optimal utilisation of TLs and/or black-start (BS) units to crank non-BS units. Optimal power flow problems are formulated for utilisation of TLs. They offer solutions for neighbouring TSOs to determine the extent to which TLs can be used without violating constraints. The proposed decision support tool is tested with IEEE 39-bus system.

Large-scale power system controlled islanding based on Backward Elimination Method and Primary Maximum Expansion Areas considering static voltage stability

This paper proposes a novel approach for separation of bulk power system into several stable subsystems following a severe disturbance. An interconnected power system may become unstable due to wide area contingency when it is operated close to the stability boundaries as a result of increased demand, power industry restructuring and competition in the deregulated electricity markets. Meanwhile, large-scale power system controlled splitting is the last resort to prevent catastrophic cascading outages and wide area blackout. The proposed method of this paper reduces the huge initial search space of the islanding strategy to only interface boundary network by clustering the coherent generators and simplifying the network graph. Then, Backward Elimination Method (BEM) based on Primary Maximum Expansion Areas (PMEAs) has been proposed to generate all proper islanding scenarios in the simplified graph. The Newton–Raphson power flow method and Q–V modal analysis have been used to evaluate the steady state stability of the islands in each generated scenario. Binary Imperialistic Competitive Algorithm (BICA) has then been applied to minimize total load-generation mismatch considering no-isolated bus, voltage permitted range and static voltage stability constraints. Comprehensive discussions have been provided using the simulations on NPCC 68-bus test system. The results demonstrate the speed, effectiveness and capability of the proposed strategy to generate fast feasible splitting solutions considering static and dynamic stability.

Two-stage stochastic optimal islanding operations under severe multiple contingencies in power grids

Due to the catastrophic consequences of rolling blackouts, there is an increasing concern with the security and stability of modern power grids. Power grid islanding, as an emergency control operation method, can divide power grids into several self-sufficient islands and can then avoid wide-area blackouts. In this paper, we present a two-stage stochastic programming model to divide the power grids into self-sufficient islands before any multiple failures happen, and optimize islanding operations plan under severe multiple contingencies that lead to extreme situations where rolling blackouts may occur. Line switching and controlled load shedding are main tools for islanding, and the expected penalty for load shedding cost is minimized in consideration of contingencies with certain probabilities to happen. The presented model can give the system operator an efficient and optimal way to properly respond to outages. Several numerical experiments are performed on IEEE test cases to show the effectiveness of the proposed islanding operations method.

Two-Step Spectral Clustering Controlled Islanding Algorithm

<?Pub Dtl=""?> Controlled islanding is an active and effective way of avoiding catastrophic wide area blackouts. It is usually considered as a constrained combinatorial optimization problem. However, the combinatorial explosion of the solution space that occurs for large power systems increases the complexity of solving it. This paper proposes a two-step controlled islanding algorithm that uses spectral clustering to find a suitable islanding solution for preventing the initiation of wide area blackouts by un-damped electromechanical oscillations. The objective function used in this controlled islanding algorithm is the minimal power-flow disruption. The sole constraint applied to this solution is related to generator coherency. In the first step of the algorithm, the generator nodes are grouped using normalized spectral clustering, based on their dynamic models, to produce groups of coherent generators. In the second step of the algorithm, the islanding solution that provides the minimum power-flow disruption while satisfying the constraint of coherent generator groups is determined by grouping all nodes using constrained spectral clustering. Simulation results, obtained using the IEEE 9-, 39-, and 118-bus test systems, show that the proposed algorithm is computationally efficient when solving the controlled islanding problem, particularly in the case of a large power system.

A Comparative Study on Frequency Estimation Methods

In this paper, a comparative study on the frequency estimation methods using IRDWT (Improved Recursive Discrete Wavelet Transform), FRDWT(Fast Recursive Discrete Wavelet Transform), and GCDFT(Gain Compensator Discrete Fourier Transform) is presented. The 345[kV] power system modeling data of the Republic of Korea by EMTP-RV is used to evaluate the performance of the proposed two kinds of RDWT(IRDWT and FRDWT) and GCDFT. The simulation results show that the frequency estimation technique based on FRDWT could be the optimal frequency measurement method, and thus can be applied to FDR(Fault Disturbance Recorder) for wide-area blackout protection or frequency measurement apparatus.

A frequency monitoring system development for wide-area power grid protection

There have been recent research activities on GPS-based FNET to prevent wide-area blackouts by monitoring frequency deviation. This paper introduces a system for monitoring regional frequencies in power grid developed as an advanced research project for intelligent wide-area protective relaying in South Korea. The system is implemented by modeling an actual 345 kV transmission system of South Korea using EMTP-RV and by measuring voltages and currents at five regions. The frequencies are estimated using our previous proposed method, a frequency estimation algorithm with gain compensation. The monitoring system implemented is used to examine frequency propagation effects and various fault conditions. Simulation results showed that the proposed system could achieve good performance regarding the frequency monitoring under both steady states and dynamic fault conditions.