Leveraging Artificial Intelligence for Enhancing Power Grid Resilience to Extreme Weather Events: Applications and Challenges

Many electricity customers in rural areas served by single radial transmission or sub-transmission line have been experiencing long duration of power outages. The need for reliable power supply, grid resiliency and rapid restoration is pushing utilities to take a renewed look at the operation of these rural feeders. One possible solution is to operate rural communities as a microgrid when the power supply line is out of service due to a permanent fault or for any other reason. The work presented in this paper examines technical and economic feasibility of operation of the two such remote communities in the rural areas of New York in a microgrid mode. The focus of this paper is to present high-level results of the case study and also to provide recommendations for utilities to evaluate the feasibility of microgrid operation as a potential solution to improve their power supply reliability and grid resiliency.

The power grid is one of the largest man-made critical infrastructures. It has been designed to distribute electric power from generating units to residential, commercial and industrial end-users. Due to the continuous increasing of electrical penetration, the availability and reliability of network is of paramount importance. In addition, the continuous increasing of renewable generators posed a further challenges to the stability of the network due to their dependencies on environmental changes, which are drifting weather scenarios towards extremes. Hence, resilience is becoming a major concern for the future power grid. In order to respond promptly to those important changes, the resilience of the such critical infrastructure has to be augmented. This can only be achieved with the availability of robust computational models that allow to design a better network, robustly validated and updated the results. Ideally, a computational framework for the assessment of power grid resilience should capture all the relevant physical interactions between components, subsystems and the system as a whole. Furthermore, uncertain and heterogeneous environmental factors have to be accounted for and their effect on safety-related metrics explicitly modelled and quantified. This is necessary to reveal power grid risks, hazards and identity situation for which an immediate safety and resilience enhancement is necessary. In this thesis, the existing power grid safety-related concepts (i.e. reliability, risk, vulnerability and resilience) and ancillary uncertainty quantification methods are analysed. The major weakness in existing quantification frameworks has been identified as the way a lack of data required by the frameworks and the treatment of such imprecise information. To overcome this limitation, a novel and robust methods for the uncertainty quantification in power grid safety-critical evaluations has been developed. The main contributions of this dissertation are a set of novel tools for the assessment of power grid reliability, vulnerability and resilience and accounting for a rigorous treatment of lack of data uncertainty. These methods have a limited need for artificial model assumptions, which might alter the quality of the available information and, with it, the validity of safety-critical decisions. One of the key elements for a resilient grid is the system ability to learn from past events, improving the grid structure, operations and policies. For this reason, a Reinforcement Learning framework for optimal decision-making under uncertainty has been investigated. This allows to equip the systems with learning capabilities, which is a fundamental component of the resilience concept, and it optimizes operation and maintenance decisions. The developed frameworks can be used to investigate the effect of threatening scenarios (such as extreme weather conditions, multiple contingencies and cascading events) on the grid safety performance. The validity of the approaches has been tested on scaled-down power grids and prognostic health management as well on realistic models of existing systems (e.g. the IEEE reliability test system). These tools provide a valuable contribution to the research community and industrial practitioners as they can help to discern whether the available information suffices to answer a reliability, vulnerability or resilience related question. If the information is limited and additional data has to be gathered, the method informs the decision-maker with the most relevant and sensitive factors, i.e. a basic indication on where to start collecting data so that an expected reduction in uncertainty is maximised.

Radley Horton,1,a Daniel Bader,1,a Yochanan Kushnir,2 Christopher Little,3 Reginald Blake,4 and Cynthia Rosenzweig5 1Columbia University Center for Climate Systems Research, New York, NY. 2Ocean and Climate Physics Department, Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY. 3Atmospheric and Environmental Research, Lexington, MA. 4Physics Department, New York City College of Technology, CUNY, Brooklyn, NY. 5Climate Impacts Group, NASA Goddard Institute for Space Studies; Center for Climate Systems Research, Columbia University Earth Institute, New York, NY

Power systems are exceedingly faced with extreme events such as natural disasters and deliberate attacks. In comparison, the underground natural gas system is considered less vulnerable to such extreme events. We consider that the overhead power grid can be hardened by replacing segments of electric power grid with underground natural gas pipelines as an energy transportation system to countereffect extreme events which can damage interdependent infrastructures severely. In this paper, an integrated electricity and natural gas transportation system planning algorithm is proposed for enhancing the power grid resilience in extreme conditions. A variable uncertainty set is developed to describe the interactions among power grid expansion states and extreme events. The proposed planning problem is formulated as a two-stage robust optimization problem. First, the influence of extreme events representing natural disasters is described by the proposed variable uncertainty set and the proposed robust model for the integrated planning is solved with the grid resilience represented by a set of constraints. Second, the investment decisions are evaluated iteratively using the conditional events. The integrated electricity and natural gas planning options are analyzed using the modified IEEE-RTS 1979 for enhancing the power grid resilience. The numerical results point out that the proposed integrated planning is an effective approach to improving the power grid resilience.

Due to increasing threats on power systems from various extreme events such as adverse weather and cyber/physical attacks, research on power grid resilience is recently gaining a substantial traction. In this study, we evaluate the transmission grid resilience using the local topological summaries derived under a framework of topological data analysis (TDA) and more conventional power system reliability metrics. The dynamics of persistent topological features after an extreme event are examined to evaluate the impact on the underlying network structure. In addition, a framework based on an optimal power flow model is developed to investigate power system reliability metrics under extreme events. The developed methods are applied to a synthetic power system that is built on the footprint of the Texas power system. By comparing the TDA summaries with the power system reliability metrics, our findings show that local topological summaries can successfully reflect changes in the grid resilience.

Hurricane Sandy and other recent extreme weather events, which have caused significant service interruptions including power outages, have revealed that our power grid is not sufficiently planned and operated to be resilient to large-scale events. Power grid resilience is a generic term that covers many aspects of power grid planning and operation. A resilient power grid should be capable of executing preventive measures, mitigating the impact of extreme events, responding optimally by automated control procedures, reducing the time needed to restore service to consumers, and potentially leading to significant economic savings. In addition, due to the growing usage of information and communication technologies (ICT), power grids are being increasingly exposed to cyber attacks. The recent blackout in Ukraine serves as a good example to illustrate the severe consequence that cyber attacks on the power grid can entail. Extensive reports and media coverage indicate the urgency to improve our power grid resilience. Maintaining resilience of the power grid is crucial to our nation’s energy security and sustainability.

The increase in global natural disasters, emergencies, and terrorist attacks brings great challenges to the flexible operation of power grids. The research on the flexibility of power grids becomes more and more important. At the same time, the rapid development of renewable energy and smart grids also provides new opportunities for research of flexible power grids. Based on the dynamic inoperability input–output model (DIIM), this study creatively uses the node transmission power to represent the relationship between nodes in the power grid and then studies the resilience measurement of the power grid. First, the significance of studying power grid resilience is discussed, and the absorption and recovery capacity of power grid resilience is described. Second, the power grid resilience measurement model based on the DIIM is established, and a theoretical analysis is carried out. Using the characteristics of inoperability, the power grid resilience is measured from the time aspect, and the characteristics of the model are analyzed at the same time. On the basis of theoretical analysis, this study also provides a control strategy to improve the grid resilience, which has good convenience and applicability. Finally, the theoretical analysis method is applied to a power grid example, and the correctness and effectiveness are verified. The measurement model can intuitively display the resilience characteristics of power grids and provide theoretical support for management and emergency response.

Extreme weather events are increasing in frequency and intensity as a result of modern climate change. During winter, species may be especially vulnerable to extreme weather as they are surviving on scarce resources and living at the edge of their thermal limits. We compiled data from eBird, a global citizen science initiative, to examine how 41 eastern North American birds shifted their occurrence and abundance patterns immediately following two recent extreme weather events each affecting > 2 million km 2 , the intrusion of a polar vortex and a winter heat wave. eBird data is continuously collected at high spatiotemporal resolution across large spatial extents, allowing us to compare species' responses immediately before and after these extreme events with trends in other winters across geographic scales. Overall, we found that birds responded differently to each extreme weather event. Bird occurrence rates did not change following the polar vortex, but where species occurred, population density was temporarily reduced, suggesting reductions in number of individuals driven by decreases in behavioral activity or temporary movement out of the area. However, birds demonstrated widespread increases in occurrence and increases in density and number of individuals where they occurred for at least 20 days after the heat wave, hinting at longer‐term range changes. Smaller‐bodied, warm‐adapted passerines tended to be most sensitive to extreme weather and responded most negatively to the polar vortex and most positively to the heat wave, while larger‐bodied, cold‐adapted waterbirds expressed only mild responses to either event. Thus, certain species may be exceptionally sensitive to extreme weather events while others are less sensitive. As climate change progresses and climatic variability increases, researchers and managers must better quantify the broad‐scale sensitivity of different species to multiple types of extreme weather events.

Electrical energy is a vital part of modern life, and expectations for grid resilience to allow a continuous and reliable energy supply has tremendously increased even during adverse events (e.g. Ukraine cyberattack, Hurricane Maria). The global pandemic COVID‐19 has raised the electric energy reliability risk due to potential workforce disruptions, supply chain interruptions, and increased possible cybersecurity threats. Additionally, the pandemic introduces a significant degree of uncertainty to the grid operation in the presence of other challenges including aging power grids, high proliferation of distributed generation, market mechanism, and active distribution network. This situation increases the need for measures for the resiliency of power grids to mitigate the impact of the pandemic as well as simultaneous extreme events including cyberattacks and adverse weather events. Solutions to manage such an adverse scenario will be multi‐fold: (a) emergency planning and organisational support, (b) following safety protocol, (c) utilising enhanced automation and sensing for situational awareness, and (d) integration of advanced technologies and data points for ML‐driven enhanced decision support. Enhanced digitalisation and automation resulted in better network visibility at various levels, including generation, transmission, and distribution. These data or information can be employed to take advantage of advanced machine learning techniques for automation and increased power grid resilience. In this paper, the resilience of power grids in the face of pandemics is explored and various machine learning tools that can be helpful to augment human operators are discused by: (a) reviewing the impact of COVID‐19 on power grid operations and actions taken by operators/organisations to minimise the impact of COVID‐19, and (b) presenting recently developed tools and concepts of machine learning and artificial intelligence that can be applied to increase the resiliency of power systems in normal and extreme scenarios such as the COVID‐19 pandemic.

The 2021 China report of the Lancet Countdown on health and climate change: seizing the window of opportunity

Extreme weather events are one of the main causes of large-scale power outages in distribution systems. The changing climate has led to an increase in the frequency and severity of these events, which, if not mitigated, are expected to lead to more instances of widespread outages and the severe societal and economic damages that ensue. Protecting the power grid against such events, which are high impact yet low frequency, requires a paradigm shift in grid design practices. In recent years, many researchers have focused on the resilience of the power grid against extreme weather events by proposing various grid hardening and/or redundancy solutions. The goal of this paper is to provide a survey of the literature related to the infrastructural resilience of the power grid against extreme events. Currently, no standard definitions or metrics exist for power grid resilience, and researchers adopt various models for quantifying and assessing it. Hence, a review of the most commonly used definitions and metrics for resilience is provided first, with a discussion of their advantages and disadvantages. Next, the paper presents an extensive and critical review of the solution methodologies proposed in the literature for improving the infrastructural resilience of the power grid. The shortcomings of the current solution methods and gaps in research are identified, followed by a discussion of the future directions in research.

Legacy and shockwaves: A spatial analysis of strengthening resilience of the power grid in Connecticut

There has been a growing penetration of renewable energy sources (RES) into power grids in recent years. As extreme weather events and cyber-attacks frequently occur, grid resilience has been an important issue. Vehicle-to-Vehicle (V2V) energy trading has great potential to improve the stability and reliability of resilient power grids integrated with high-penetrated RES. V2V energy trading is a distributed peer-to-peer (P2P) application. As a decentralized distributed ledger technology, blockchain is an ideal platform for V2V energy trading. The consensus mechanism of blockchain determines whether the V2V energy trading blockchain (ETB) can improve grid resilience. However, most studies in the ETB currently utilize conventional consensus mechanisms. Due to their substantial computational requirements and communication overhead, these consensus algorithms are not well-suited for real-time service applications like energy trading. We propose a novel BAC-SDS consensus specifically for V2V ETB, thus enabling resilient grids to maximise the use of renewable energy. We propose an Electric Vehicle (EV) leader election based on cryptography and adopt the sharding technique to enhance the system's scalability. Furthermore, our approach ensures the secure transfer of energy and value, contingent upon the condition that all EVs exhibit reasonable behavior and retain the proofs they possess. We implement the V2V ETB on Hyperledger Fabric. The experiments demonstrate (1) that V2V ETB significantly enhances the resilience of the power grid compared to traditional centralized trading models and (2) the consensus mechanism proposed in this article is better suited for V2V energy trading than existing mechanisms, exhibiting superior performance in terms of security, throughput, and scalability, thus further enhancing the resilience of the power grid.

Recent severe power outages caused by extreme weather events have highlighted the importance and urgency of improving the resilience of the electric power grid. Grid resilience is increasingly critical since the number of outages caused by severe weather is expected to rise as climate change increases the frequency and intensity of hurricanes, blizzards, floods, and other extreme weather events. As the distribution grids may still remain vulnerable to natural disasters to a certain extent, the power industry has focused on methods of restoring distribution systems quickly after disasters. However, current distribution grid restoration practice is based on the predetermined priorities from previous experiences, which is not adaptive with respect to the status of power grid damage under evolving weather events and available restoration capabilities and resources, and thus tends to be inefficient and suboptimal. In addition, lack of situational awareness of distribution grids poses great challenges to power system operators, which largely delays the restoration process and incurs large economic costs to customers.

In recent years, the challenge of ice accumulation on power lines in southern China has been increasingly prominent due to frequent occurrences of extreme cold weather conditions. During the ice-melting process of the 500kV lines under the Southern Power Grid jurisdiction, there have been several instances of fuse blowing at the junctions between tower ground wires and NY-type clamps, particularly when the contact resistance between the downlead conductors and tension-resistant clamps is excessively high. This leads to a rapid temperature increase in the joint area due to Joule heating effects, resulting in interruptions in the ice-melting process and an expansion of ice formation along the transmission lines, posing a threat to the operational safety of tower equipment. Addressing the critical issues affecting the anti-icing capability of the power grid, and considering the high current intensity, concentrated operating time, and demanding conductive performance of the ice-melting process, a structurally optimized, superbly conductive, and conveniently installable TY-Type Clamp has been devised to replace the existing downlead line clamps and form an ice-melting line. By optimizing contact pressure and conductive area, this clamp markedly reduces contact resistance, effectively suppressing localized overheating during ice-melting, achieving the objective of swift repairs for disconnected downlead conductors on overhead ground wires in ice-melting lines. It furnishes a reliable technical solution for rapidly restoring ice-melting operations and preventing accidents from escalating during ice accumulation periods, holding vital engineering significance in enhancing the power grid's resilience against ice disasters.