Shinichi Imai, Damir Novosel, Daniel Karlsson, Alexander Apostolov
Power system blackouts result in complete interruption of the electricity supply to all consumers in a large area. Systems need to be planned and engineered in a way that minimizes exposure to and prevents cascading blackouts. Blackouts are caused by a sequence of cascading outages caused by a combination of multiple low-probability events (e.g., a transmission line sagging into a tree, hidden failures in equipment protection, the loss of multiple generation units because of a weather event, and aging equipment failure) occurring in an unanticipated or unintended sequence. The likelihood for power system disturbances to escalate into a large-scale cascading outage increases when the grid is already under stress. This stress can be caused by lower operational margins, overloaded equipment, and other factors, which are discussed in detail in this article.
Since the electric power industry is rapidly changing, a resilient electrical grid provides a foundation for achieving societal goals by enabling the integration of renewable energy resources, electrical storage, and electrification. Those changes require new ways of planning, operating, maintaining, and regulating the grid. The grid has been experiencing an increased impact on resilient and reliable operation due to 1) natural disasters and pandemic events, 2) human-made cyber- and physical attacks, and 3) events resulting from system design, aging, and inadequate human response. This article focuses on the third. Recent examples show that system operational margins were significantly reduced due to generation capacity limits and that inverter-based generation was unnecessarily tripping due to faults on the transmission system. It is apparent that unless adequate and fast actions are taken, unexpected consequences can occur across the electrical grid, leading to severe cascading failures.
The integration of renewable energy and new technologies is creating both opportunities and challenges for power system resilience and reliability. The penetration of large-scale power electronic inverter-based resources (IBRs) has caused faster and unpredictable dynamic changes of the power system parameters, which may lead to higher risks of cascading outages. Those systems experience issues, such as low fault currents and low system inertia, that may negatively affect grid performance unless modifications in power system planning, operation, and control are made to address the changes of power system parameters. At the same time, those technologies offer new solutions to improve reliability and resilience during wide-scale outages. For example, renewable distributed energy resources (DERs) and storage, in various forms, including microgrids, have demonstrated an ability to serve the load and support restoration for various system events. Preventing cascading outages requires significant electrical power system improvements, such as better visibility, accurate system and equipment models and tools, automated schemes for arresting disturbance propagation and speed up restoration, and other technological advances.
Blackouts often occur in stressed electrical power systems with low operational margins. Those systems need to adapt to changes in the use of electrical energy. For example, the grid was not originally designed for deregulated markets, created in the 1990s, that included transferring large amounts of electrical power across interconnections. Since both the grid and the operators were not able to manage fast-developing disturbances, the result was a global increase in major blackouts. Because the power industry is addressing decarbonization needs with a large integration of renewable generation and electrification and since the system is significantly affected by increased weather disruptions, it is necessary to redesign the transmission and distribution (T&D) grid to address those changes. After blackouts in Europe and North America in 2003, our industry successfully addressed the grid needs by making major investments in infrastructure upgrades, innovative hardware and software solutions, and processes. Since it is better to address major issues before they occur, our industry and society need to take actions to address present and upcoming changes in how electrical power is produced and delivered to prevent future blackouts. Some areas have already experienced brownouts and planned shutdowns to avoid major blackouts, so a major focus needs to be put on managing, planning, and designing the grid to address various operational scenarios for blackout prevention and restoration to secure reliable and resilient power delivery.
Since systems with large penetrations of renewable energy resources and storage exhibit more dynamic behavior than systems with conventional generation, it is particularly important to design systems that can deploy automated actions to address those scenarios. Generally, disturbance propagation involves a combination of several phenomena:
Voltage instability/collapse: These problems usually occur when the power transfer is increased, resulting in inadequate voltage support. This is often a result of not installing needed transmission lines and voltage support devices when generation resources are far away from load centers. Voltage stability could be classified as short-term (transient) and long-term.
The key to making the grid more resilient is to effectively address preconditions that cause grid stress, such as the following:
Reducing disturbance propagation and enabling faster restoration require the deployment of monitoring, control, and protection devices; software tools; telecommunication infrastructure; and hands-on lifecycle training programs that help improve the management of the operational aspect of the grid. However, such actions do not replace the investment needed in the grid infrastructure through asset management.
Furthermore, it is quite difficult for operators to prevent cascading during a fast-developing dynamic disturbance, as they need to coordinate with neighboring system operators, verify equipment ratings and statuses, and implement corrective measures. Automated actions are generally the most effective response to stabilize the system.
With the increased penetration of distributed renewable energy resources, monitoring weather conditions becomes even more critical due to weather’s impact on the output of the DERs. The weather also can play a role in the loading of the system, which will increase or decrease because of changes in heating and air conditioning loads.
The trend of increasing IBRs is imposing emerging challenges to reliable system planning and operation. With the increased penetration of IBRs, it is critical to ensure that IBRs can provide essential reliability services to the electric power system. A summary of key concerns with IBR integration and associated system aspects is provided in Table 1.
Table 1. The IBR integration issues.
Traditionally, the power grid has relied on stored mechanical energy in large generators’ rotating masses (inertia) and other tools and technologies, such as central control coordination, to maintain grid stability and balance during moderate disturbances. As more DERs using IBRs without inertia are deployed, there will be a growing imbalance between generation and load, resulting in the need to review underfrequency load shedding (UFLS) schemes. The grid’s architecture and the associated controls technology can transition to provide similar functionalities even as traditional forms of generation are replaced by more carbon-free generation. While conventional synchronous generators have operating reserve margins and “naturally” compensate for imbalances, IBRs can better react immediately and much faster than conventional generation in balancing load and generation through improved management and coordination with system needs, which are achievable with present technology. IBRs with smart inverters need to have reserve margins and be controlled at the system level to be effective in balancing generation and load, which requires both coordination and appropriate market mechanisms.
IBR resources do not provide any significant increase in current during short circuit events; rather, they provide either no change in current or only a very nominal amount during the short circuit events. As traditional rotating synchronous generators are retired and replaced with IBRs, it is expected that the system will experience a fundamental change in short circuit behaviors across all levels of the grid, specifically, lowering currents and the strength of short circuits. The existing protection and control systems will have major challenges in detecting faults with a high penetration of renewables, creating possibilities for incorrect protection operations leading to blackouts. Solutions are either grid equipment that provides fault current or new protection and control schemes. To remain in service following a short circuit fault, DERs need to be able to withstand the voltage drop until the fault has been cleared by the protection and the circuit breaker. The increased fault clearing times that are caused by the in-feed effect of a distributed generator may not be acceptable to customers with sensitive loads. The voltage sag is experienced not only by users on the faulted feeder but also on the adjacent feeders connected to the same distribution system. Considering that the protection operating time depends on the type of protection, location of the fault, type of fault, and fault parameters, these factors need to be analyzed to identify the requirements for improvement in the fault clearing time.
The impact of voltage sags and swells on DERs depends on two characteristics. The first characteristic of a voltage sag—the depth—is a function of the type of fault, fault location, and system configuration. Single phase-to-ground faults lead to voltage sag in the faulted phase and to voltage swell in the healthy phases. The level of voltage increase is also affected by the grounding of the interface transformer and should also be taken into consideration. The second characteristic is the duration of the voltage sag, which depends on the fault clearing time. These two characteristics of the fault have an impact on the ride-through capability of the DER. Figure 1 gives an example of a ride-through characteristic.
The first characteristic is something that we cannot control but that we must study to be able to predict and estimate the effects of different faults on system stability. The second characteristic of the voltage sag—duration—is the parameter that we can control by properly applying the advanced features of multifunctional protection relays and state-of-the-art communication technology.
Another impact of inverter-based DERs on the stability of the electric power grid has to do with the challenges of their modeling in transient simulation and dynamic stability programs. The algorithms used in the inverter controllers are not standardized and, in many cases, are considered company trade secrets. As a result, the studies performed as part of the design of system integrity protection schemes (SIPSs) may not accurately represent the real behavior of the system during conditions that may lead to local and wide-area disturbances and increase the probability for blackouts. That is why it is necessary for the industry to identify methods for addressing these issues.
Widespread electric outages are symptoms of inadequate grid management strategies. The analysis of four blackouts in Figures 2 and 3 reveals common threads among them. The figures describe the chain of events that led to four blackouts, the duration of those blackouts, and the time to restore the systems. The conclusion is that it is necessary to learn from the past and implement proven methods to mitigate the impact. Further analysis shows that although it is not possible to avoid multiple contingency-initiated outages, blackout propagation could and should be arrested, and the restoration time could and should be reduced through appropriate black start procedures.
The following are the main conclusions from analyzing the chain of events:
In general, the events encompassed practically an infinite number of operating contingencies differing from the expectations of the system designers.
South Australia Blackout in 2016
Violent storms caused damage to transmission equipment and resulted in consequent multiple line faults and tripping by protection devices. This disruption resulted in the reduced output of one wind farm and the shutdown of nine wind farms, due to the ride-through settings of inverters, which tripped after repeated voltage deviations, causing a sustained reduction in power of 456 MW. Those events overloaded the interconnections to Victoria, which subsequently tripped, causing cascading outages, including the tripping of the remaining generation and the loss of supply to the entirety of South Australia. The Australian Energy Market Operator identified issues with the inverter settings of the wind farms that tripped prematurely, as others stayed online until the system went into blackout.
Hokkaido Blackout in Japan in 2018
As a result of an earthquake, N-4 contingencies caused the complete blackout of one of the regions in Japan. The shutdown of triple units of a thermal power plant (an N-3 contingency) and the resulting loss of four lines (N-4) initiated system separation, followed by the tripping of some hydropower plants. An emergency power transfer was initiated to recover the frequency drop by utilizing a high-voltage dc (HVdc) link with an adjacent region, but it was not sufficient to prevent the tripping of the remaining major power plant and the complete blackout of the region. Short-term preventive actions included adding UFLS of 350 MW.
Puerto Rico Blackout in 2022
This blackout, affecting close to 600,000 people, is an example of inadequate equipment maintenance and investment in the grid. A single phase-to-ground fault of 115 kV, caused by overgrown vegetation, was cleared by protection. However, one of the circuit breakers failed to operate due to a dc power supply failure. In addition, the breaker failure scheme also failed to operate due to dc power supply problems, resulting in cascading outages across the system and system separation. The extent of the outage was exacerbated by several additional failures to operate, incorrect operations, and delayed operations.
Additional disturbances, which resulted in major customer outages but did not result in system-wide blackouts, are described in the following.
United Kingdom Electrical Power Disturbance in 2019
A number of DERs connected to the distribution network were disconnected automatically immediately following a lightning strike. Two large wind generators, the Little Barford steam turbine (244 MW) and Hornsea 1 offshore wind farm (737 MW), were unable to continue providing power to the system. As a result of this combined loss of generation, the system frequency fell rapidly, causing a larger volume of DERs to disconnect. Load shedding was therefore triggered to contain the power outage.
European System Split in 2019
The system experienced high active power flow from east to west (warm weather in the southeast and a cold spell in the northwest, causing high demand), resulting in low system margins. A planned outage on a 400-kV system resulted in a busbar coupler overload protection trip, followed by transformer and transmission line overload protection tripping, which then caused system transient instability and the cascading tripping of a number of 400- and 220-kV transmission lines. The system separated into two areas, one low frequency and the other high frequency. The frequency in the west system was arrested at 49.74 Hz by disconnecting 1.7 GW of interruptible load and recovered through follow-on automated activation of 504 MW of reserve generation. The frequency rise in the east was arrested by disconnecting 975 MW of generation. The two systems were reconnected after about an hour, including the tripped load.
Texas Electrical Power Disturbance in 2021
Due to a cold spell in Texas, USA, 1.365 million customers were affected. Rotating outages were initiated for ∼3.9 million customers (30%). This was primarily a generation issue, as approximately 48.6% of all types of generation were out at the highest point, due to nonwinterized gas wells and gas pipelines, the variety of generators, interruptible gas supply contracts to the generators, and insufficient coordination between the electric grid and gas facilities. The system frequency was below 59.4 Hz for 4 min, 23 s, but the automated UFLS set point of 59.3 Hz was not reached, and the system was not blacked out. Additional generators would have tripped if the frequency had been below 59.4 Hz for 9 min. Some distribution systems were damaged, including the loss of heating systems and outages of municipal water treatment systems, requiring the evacuation of some hospitals. Approximately 2,300 transformers and 22,000 fuses were replaced, as well. An unprecedented increase in electric power market prices was experienced during the event.
Several activities are necessary to prevent blackouts and minimize their effect. The process should start with understanding the power system characteristics. The next step is to design and operate the power system so that blackouts can be avoided to a reasonable extent. Measures to prevent blackouts could be grouped into the following major categories:
Reliability standards and grid investments, including aging infrastructure and control technology:
• System planning and operations:
• Monitoring, protection, and control:
Modern communication and computer technologies facilitate the deployment of an overall system-wide protection and control philosophy, as it is possible to tie all the monitoring, control, and protection devices together through an information network. The key to a successful solution is fast detection, fast and powerful control devices, high-speed reliable communication infrastructures, and smart algorithms—in other words, WAMPAC. Phasor measurement units (PMUs) are very well suited as transducers for the input data since they provide properly timed synchronized measurements of currents and voltages directly from the instrument transformers. WAMPAC systems can help in better congestion tracking, visualization, information sharing over a wide region, and so on. An example is wide-area voltage stability management. PMUs have been incorporated in various voltage stability monitoring and contingency analysis tools, such as EMS and model-based voltage stability assessment (VSA) tools for slower-developing disturbances. However, for dynamic voltage instability, fast measurement-based voltage instability detection tools using PMUs could be a good complement to VSA tools. Since those tools do not require a complex power system model, they are not as sensitive to model inaccuracies, and detection can occur at a subsecond rate. They are also suitable for implementation in SIPSs and, in the simplest case, instead of undervoltage protection. They are able to identify unstable cases even in the case of high voltage (for which undervoltage load shedding relays would not operate) and stable cases in the presence of very low voltages (for which an undervoltage relay should not operate).
Automated actions, such as SIPSs, are important to avoid fast disturbance propagation. SIPSs include accurate and reliable monitoring systems with faster detection and wider communication capabilities and play a key role in the development of a comprehensive system defense plan for reducing the risk of large system-wide disturbances and preventing blackouts. Time frames for the overall grid management, including SIPSs and WAMPAC, are given in Figure 4. The following section addresses the deployment of SIPSs, often called special protection schemes (SPSs) and remedial action schemes.
SIPSs are aimed to preserve power system integrity if the system is in a transition toward instability. If no remedial action is taken, the system will go unstable, and a blackout will follow. Such remedial actions can, for example, be load shedding, based on voltage or frequency criteria, and generation curtailment, based on line trips and line corridor overloads. Since the trend in power system planning has been toward tight operating margins with less redundancy and more complex systems due to the increased integration of DERs and storage, SIPSs are beginning to play an even more important role.
SIPSs are categorized as event based, parameter based, response based, and a combination of the three. Event-based SIPSs act directly on a power system event, such as a breaker trip, and take remedial actions, e.g., generation curtailment in the sending end, due to a line trip in a transmission corridor. Event-based SIPSs are fast, but the drawback is that every “event” to be handled—it must be identified beforehand. Parameter-based SIPSs measure variables for which a significant change confirms the occurrence of a critical event. Response-based SIPSs act on the system response in terms of voltage, frequency, power flow, and so on, regardless of the actual event causing the response. Response-based protection schemes may tend to be slower, waiting for the system response, but are more general.
In power systems with high penetrations of IBRs, event-based and parameter-based schemes can have challenges associated with security (the ability to always operate when required) and dependability (the ability not to operate when not required). Since those schemes rely on offline network analysis for decision-making logics and to determine parameters, dynamic characteristics of IBRs (including ride-through capabilities) and dynamic load performance, such as induction motor stalling, affect their performance. Response-based schemes monitor the system response during disturbances and incorporate a closed-loop process to react to the actual system conditions. However, more analysis and deployment of technologies, such as synchronized measurements, are required to achieve the expected performance of SIPSs against the dynamic response of systems with high penetrations of IBRs.
Specific countermeasures should be designed for each disturbance category, as identified previously, e.g., Figure 5. For each category and power system condition, it is then possible to design proper SIPSs. All control systems include inputs, output, and a control algorithm, or decision-making process, in a feedback loop to the controlled system, as indicated in Figure 5.
The power system is exposed to disturbances of different kinds that a SIPS is intended to counteract. On the input side, we find continuous state variables from all relevant parts of the system, such as voltage magnitude and phase angles, frequency, and active and reactive power, as well as binary condition variables, such as switching states of breakers. On the action side, we find voltage boosting, fast valving, load shedding, reactive power support activation, generation curtailment, and so on. For the case of redundancy, each action should preferably have its own protection scheme. The decision-making algorithm could be more or less sophisticated, perhaps including only the communication of binary values, such as “low voltage,” “very low voltage,” “extremely low voltage,” and so on.
SIPSs add an extra layer of complexity to power system operation, and challenges regarding the use of system protection are mainly connected to the ability to maintain an overview of the active SIPSs in the system. This situation must be resolved by custom-made systems and good training of the operators. When designing SIPSs, it is very important to consider what could happen if a SIPS has a malfunction and if it is used in the wrong conditions. An allocation of a certain SIPS also affects other SIPSs. A more detailed overview of SIPS design and operation is presented in Figure 6.
SIPSs can be simple and complex, with different numbers of levels in the hierarchy, depending on the complexity of the system. Figure 7 shows a multilevel SIPS that uses multifunctional protection intelligent electronic devices (IEDs) as the devices at the monitoring and execution levels of the system. A wide-area protection system requires different types of communications:
All the preceding communications interfaces may be based on different protocols and use different types of communications links.
International Electrotechnical Commission (IEC) standard 61850 could play an increasingly important role due to the significant benefits that high-speed peer-to-peer messages play in the implementation of different functional elements of the scheme. One of the key components that gives it this role is the generic object-oriented substation event (GOOSE) message that supports peer-to-peer communications among multifunctional IEDs, which meet the high-speed performance requirements of protection and automation applications. It was originally designed as a nonroutable message used within the substation over the local area network. The success of using GOOSE messages for substation protection applications makes the messages attractive for use among IEDs in wide-area protection systems. These are applications that impose different requirements on the peer-to-peer communications.
Even though GOOSE messages have already been used in protection and automation applications outside of the substation, the fact that they are transmitted over wide-area networks makes them vulnerable to cyberattacks. To address such concerns, IEC Technical Committee 57 Working Group 10 developed the technical report IEC 61850 90-5, which defined the communications of PMUs and GOOSE messages over wide-area networks. The availability of secure peer-to-peer communications allows the use of routable GOOSE (R-GOOSE) messages in distributed wide-area protection systems, the bidirectional exchange of information among field IEDs, the wide-area protection system controller for the monitoring of the power system, and the execution of required actions to isolate faulted sections and maintain the stability of the electric power grid. The R-GOOSE described previously has multiple applications in SIPSs. It brings some significant benefits for wide-area distributed applications, especially when they are based on wireless communications technologies. The cybersecurity features defined in IEC 61850-90-5 and IEC 62351 provide a high level of security, which is a key requirement for SIPSs. R-GOOSE messages can be used for both vertical and horizontal communications in hierarchical SIPSs.
Statnett, Norway
Statnett, the Norwegian transmission system operator (TSO), uses SIPSs to improve the trading capacity, increase the transmission capacity, and increase the reliability of the power supply. SIPSs reduce the consequences from a single fault, and it is possible to increase the power flow in the system without increasing the risk for overload. The use of system protection can also postpone the need for new lines. In case of critical line trips and line overloads, Statnett uses the automatic disconnection of production and load, changes in the flow on the HVdc connections to Denmark and The Netherlands, and automatic circuit breaker trips to split networks to keep the grid energized and the power flows within set limits. The value of SIPSs in the Norwegian power system is high and results in increased revenues and decreased costs through raised transfer and trading capacity. It also gives enhanced security in the system and, in general, better use of the power resources. It is estimated that the revenues due to increased trading capacity from the use of SIPSs amounts to €25–50 million. Statnett covers the network owner, producer, and consumer expenses related to SIPSs. Figure 8 illustrates the design of a SIPS used to reduce generation in case of a transmission line trip. The yellow boxes detect a line trip, and the red box detects a line overload for a critical line; the logic then trips or reduces generation to avoid line trips due to overloads in the weak parts of the system.
The consequences of faulty activations of SIPSs can be quite serious. So far, Statnett has not been subjected to very serious events. Some smaller events have occurred. A few examples are as follows:
These events are unfortunate and have some cost, but the consequences are not very serious, after all.
In parts of the Norwegian power system not operated by Statnett, attempts have been made to realize SIPSs through programming the SCADA system. This use has proved to be a rather unstable solution, which has resulted in a number of unwanted triggers of the SPSs. Statnett will not recommend such a solution.
TEPCO, Japan
TEPCO Power Grid, one of the Japanese TSOs, has applied UFLS to avoid blackouts. As the local SIPS, the centralized logic with digital relays in the substations initiates to trip the feeders in case of detecting underfrequency and RoCoF decay. The UFLS operated as designed for the drop of massive numbers of generators following major earthquakes three times in the past 40 years, including 2011, 2021, and 2022.
Since the integration of DERs, mainly in the sub-T&D level, has recently grown significantly, the feeders carrying reverse power flow may be shed by UFLS, which induces the degradation of the expected frequency recovery by the UFLS program. Recalculation of UFLS settings is globally done using EMS information. The signals to the local SIPS from the global SIPS are sent via an IEC 61850 communication network, as detailed in Figure 9. The existing UFLS has a protection logic to detect the RoCoF by measuring the time duration between two levels of underfrequency during the frequency decay, instead of detecting the differential of the frequency with respect to time. The current logic is reliable for temporary frequency deviation during oscillation but may reset the RoCoF detection unexpectedly, as demonstrated in Figure 10. Improved logic using integral components to balance security and dependability has been studied and tested. Adaptive UFLS with a hierarchical architecture composed of global SIPSs in the control center and local SIPSs in the substations is planned to be implemented in the near future by TEPCO Power Grid.
Cenace, Ecuador
The Ecuadorian electrical power system has experienced rapid generation expansion and demand growth leading to stressed grid operations. A system analysis showed that double contingencies could cause a system collapse. The solution was to deploy a centralized SIPS to prevent the collapse of the 230-kV backbone transmission ring to avoid the partial or total collapse of the whole system. Independent system operator Cenace has implemented a centralized fully redundant system, shown in Figure 11, involving two control centers and 12 transmission company substations and 11 distribution company substations. For each precalculated contingency and based on predetermined system measurements, the system trips selected generators and loads in fewer than 200 ms.
The system was deployed in 2015 and has been expanded through multiple phases and operated multiple times. It operated three times in 2015, with estimated total savings of around US$4 million, resulting from reducing energy not supplied of almost 3,000 MWh/year. The SPS return on investment was less than one year.
Modern power systems, with large penetrations of renewable energy resources, storage, and electric vehicles, exhibit more dynamic and faster behavior, including variable system margins. Under these conditions, risks of blackouts will increase unless proactive actions are taken. The actions include investments not only in physical infrastructure and new technologies but also in advanced automation (e.g., WAMPAC and SIPSs), tools, and processes to address various system scenarios, especially scenarios that have been considered low probability but are increasingly common. To emphasize the IEEE definition of resilience that applies to blackout prevention, we need to “protect against and recover from any event that would significantly impact the grid by preparing, withstanding, and reducing magnitude and duration of the event.” Furthermore, connected power systems should be designed so that individual control systems perform correctly to limit and isolate problems before they lead to widespread cascading outages.
WAMPAC, using synchronized measurement technology, offers still-untapped opportunities to improve detecting and controlling fast-developing disturbances. Real-time high-speed state measurements, such as phase angles and rate-of-change angles, provide the base for response-based SIPSs and enable combining multiple schemes into a hierarchical one.
This article has focused on solutions to improve the response to blackouts caused by natural threats, human-made threats, and/or system design issues. Further attention should be paid to “black start” processes and tools for faster restoration in the presence of IBRs, microgrids, mobile storage, and vehicle-to-grid technologies.
D. Novosel and V. Madani, “System integrity protection schemes,” in Encyclopedia of Electrical and Electronic Power Engineering, J. García, Ed. Oxford, UK: Elsevier, 2022, pp. 132–142.
A. Derviskadic, “Analysis of the two systems splits and one inter-area oscillation within the continental European synchronous area in year 2021,” in Proc. Int. Conf. Smart Grid Synchronized Meas. Analytics (SGSMA), Split, Croatia, May 2022.
B. Chalamala et al., TR-93: Importance of T&D Grid Modernization to Mitigate Impacts from and Adapt to Climate Change. Piscataway, NJ, USA: IEEE PES, 2022.
M. V. Flores, D. E. Echeverría, R. P. Barba, and G. Argüello, “Architecture of a systemic protection system for the interconnected national system of Ecuador,” in Proc. Asia-Pacific Conf. Comput. Aided Syst. Eng., Quito, Ecuador, Jul. 2015, pp. 414–419, doi: 10.1109/APCASE.2015.80.
D. Novosel, “Transmission blackouts: Risk, causes, and mitigation,” in Contribution to Encyclopedia of Sustainability Science and Technology, M. Begovic, Ed. New York, NY, USA: Springer-Verlag, 2011, pp. 99–127.
V. Madani et al., “IEEE PSRC report on global industry experiences with system integrity protection schemes (SIPS),” IEEE Trans. Power Del., vol. 25, no. 4, pp. 2143–2155, Oct. 2010, doi: 10.1109/TPWRD.2010.2046917.
Shinichi Imai is with Takaoka Toko, Tokyo 135-0061, Japan.
Damir Novosel is with Quanta Technology, Raleigh, NC 27607 USA.
Daniel Karlsson is with DNV, SE-216 20 Malmo, Sweden.
Alexander Apostolov is with Omicron, Los Angeles, CA 90064 USA.
Digital Object Identifier 10.1109/MPE.2023.3247096
Date of current version: 19 April 2023
1540-7977/23©2023IEEE