No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Statistical Forecasting of Electric Power Restoration Times in Hurricanes and Ice Storms
Abstract
This paper introduces a new method for estimating the time at which electric power will be restored after a major storm. The method was applied for hurricanes and ice storms for three major electric power companies on the East Coast. Using an unusually large dataset that includes the companies' experiences in six hurricanes and eight ice storms, accelerated failure time models were fitted and used to predict the duration of each probable outage in a storm. By aggregating those estimated outage durations and accounting for variable outage start times, restoration curves were then estimated for each county in the companies' service areas. The method can be applied as a storm approaches, before damage assessments are available from the field, thus helping to better inform customers and the public of expected post-storm power restoration times. Results of model applications using testing data suggest they have promising predictive ability.
... Previous studies have investigated the durations of power outages after hurricanes at a regional level. Liu et al. (2007) applied an accelerated failure time (AFT) model to understand restoration time of power outage. While this approach provides useful insights for time to event data analysis, it ignores spatial clustering of restoration time for power outage. ...
... • We developed generalized accelerated failure time (GAFT)-a statistical model-to investigate the association between restoration time of power outage and a wide range of variables including hazard, built environment factors, and sociodemographic characteristics of the regions accounting for spatial dependence of observations. While power outage has been studied from the perspective of time to event data analysis (Liu et al. 2007) and considering the spatial dependence of observations (Mitsova et al. 2018), we add a new dimension by developing the GAFT model that can account both for time to event data and spatial dependence of observations. ...
... Researchers have developed models to identify the contributing factors toward power outage following a disaster. Liu et al. (2007) developed an accelerated failure time (AFT) model for determining the time required for the restoration of power outage after an extreme hazard, considering hurricane and snowstorm. Nateghi et al. (2011) compared different models such as accelerated failure time model, Cox proportional hazard model, regression trees, Bayesian additive regression trees (BART), and multivariate additive regression splines and found that BART performs the best. ...
Major disasters such as wildfire, tornado, hurricane, tropical storm, and flooding cause disruptions in infrastructure systems such as power and water supply, wastewater management, telecommunication, and transportation facilities. Disruptions in electricity infrastructure have negative impacts on sectors throughout a region, including education, medical services, financial services, and recreation. In this study, we introduced a novel approach to investigate the factors that can be associated with longer restoration time of power service after a hurricane. Considering restoration time as the dependent variable and using a comprehensive set of county-level data, we estimated a generalized accelerated failure time (GAFT) model that accounts for spatial dependence among observations for time to event data. The model fit improved by 12% after considering the effects of spatial correlation in time to event data. Using the GAFT model and Hurricane Irma’s impact on Florida as a case study, we examined: (1) differences in electric power outages and restoration rates among different types of power companies—investor-owned power companies, rural and municipal cooperatives; (2) the relationship between the duration of power outage and power system variables; and (3) the relationship between the duration of power outage and socioeconomic attributes. The findings of this study indicate that counties with a higher percentage of customers served by investor-owned electric companies and lower median household income faced power outage for a longer time. This study identified the key factors to predict restoration time of hurricane-induced power outages, allowing disaster management agencies to adopt strategies required for restoration process.
... Studies that focused on restoration assessment considering ice storms were presented in [34,35]. The work in [34] proposed accelerated failure time models to estimate outage durations and restoration times for hurricanes and ice storms. ...
... Studies that focused on restoration assessment considering ice storms were presented in [34,35]. The work in [34] proposed accelerated failure time models to estimate outage durations and restoration times for hurricanes and ice storms. In [35], the authors proposed a Monte Carlo simulationbased approach to estimate restoration times considering a moving ice storm with continuous severity levels. ...
... Short-term planning mechanisms are designed to prevent or mitigate the impacts of ice storms on power systems, when an ice storm can be forecasted accurately, and reduce the system restoration time [34]. They include decisions that can take place hours or a few days before the ice storm hits a power system. ...
Ice storms have significantly impacted power systems in many ways over the past decades, causing serious damage to power generation, transmission, and distribution assets. Such storms have led to massive and extended power blackouts that have disrupted the lives of many communities, thus raising concerns from several electricity sector stakeholders. This paper aims to provide power grid operators, engineers, researchers, and regulators a comprehensive overview and comparison of the existing short- and long-term planning and operational methods to enhance the resilience of power grids against ice storms. Moreover, the paper also discusses several challenges, needs, and opportunities for future research and development, with a special focus on planning and operation mechanisms for power transmission and distribution network operators.
... Previous studies investigated the durations of power outages after hurricanes at a regional level. (Liu et al., 2007) applied Accelerated Failure Time (AFT) model to understand restoration time of power outage. While this approach provides useful insights for time to event data analysis, it ignores spatial clustering of restoration time for power outage. ...
... Researchers have developed models to identify the contributing factors toward power outage following a disaster. Liu et al. (2007) Quiring et al. (2011) included soil characteristics and suggested that these variables can implicitly inform about the likelihood of trees being uprooted. McRoberts et al., (2018) showed that the inclusion of elevation, land cover, soil, precipitation, and vegetation characteristics improved the accuracy of previously established statistical model by 17%. ...
... Previous studies on Hurricanes Irma (Mitsova et al., 2018) and Hurricanes Bonnie, Isabell, Dennis, and Floyd (Liu et al., 2007) show that maximum sustained wind speed is positively associated with power service restoration time. Number of power plants is important to predict thunderstorm induced power outages (Kabir et al., 2019). ...
Major disasters such as wildfire, tornado, hurricane, tropical storm, flooding cause disruptions in infrastructure systems such as power outage, disruption to water supply system, wastewater management, telecommunication failures, and transportation facilities. Disruptions in electricity infrastructures has a negative impact on every sector of a region, such as education, medical services, financial, recreation. In this study, we introduce a novel approach to investigate the factors which can be associated with longer restoration time of power service after a hurricane. We consider three types of factors (hazard characteristics, built-environment characteristics, and socio-demographic factors) that might be associated with longer restoration times of power outages during a hurricane. Considering restoration time as the dependent variable and utilizing a comprehensive set of county-level data, we have estimated a Generalized Accelerated Failure Time (GAFT) that accounts for spatial dependence among observations for time to event data. Considering spatial correlation in time to event data has improved the model fit by 12%. Using GAFT model and Hurricane Irma as a case study, we examined: (1) differences in electric power outages and restoration rates among different types of power companies: investor-owned power companies, rural and municipal cooperatives; (2) the relationship between the duration of power outage and power system variables, and socioeconomic attributes. We have found that factors such as maximum sustained wind speed, percentage of customers facing power outage, percentage of customers served by investor-owned power company, median household income, and number of power plants are strongly associated with restoration time. This paper identifies the key factors in predicting the restoration time of hurricane-induced power outages.
... As the power outage data is generally not made publicly available by the utilities, the previous models are primarily calibrated to data from a few regions. For example, Liu et al. (2005Liu et al. ( , 2007Liu et al. ( , 2008 developed the outage prediction model for North and South Carolina. Guikema et al. (2014), Nateghi et al. (2014), and Shashaani et al. (2018) developed the outage prediction models for the Gulf Coast. ...
... ACS started data collection in 2010, and we have considered data from 2019. We obtained the population density as it is indicates the number of distribution poles and system components exposed to winds (Liu et al., 2007). ...
... We found precipitation and soil moisture are important for outage prediction even for linear regression, suggesting that their relevance could be even higher for non-linear regressions. We also found that population density is critical for outage prediction, which could be explained by a positive correlation between density and the density of transformers, as described in (Liu et al. (2007) Feature descriptions are shown in Table 1. In the second stage, we analyzed the correlations between the input parameters. ...
Strong hurricane winds damage power grids and cause cascading power failures. Statistical and machine learning models have been proposed to predict the extent of power disruptions due to hurricanes. Existing outage models use inputs including power system information, environmental, and demographic parameters. This paper reviews the existing power outage models, highlighting their strengths and limitations. Existing models were developed and validated with data on a few utility companies and regions, limiting the extent of their applicability across geographies and hurricane events. Instead, we train and validate these existing outage models using power outages for multiple regions and hurricanes, including Hurricanes Harvey (2017), Michael (2018), and Isaias (2020), in 1,833 cities along the U.S. coastline. The dataset includes outage data from 39 utility companies in Texas, 5 in Florida, 5 in New Jersey, and 11 in New York. We discuss the limited ability of state-of-the-art machine learning models to (1) make bounded outage predictions, (2) extrapolate predictions to high winds, and (3) account for physics-informed outage uncertainties at low and high winds. For example, we observe that existing models can predict outages as high as 25 times more than the number of customers and cannot capture well the outage variance for wind speeds over 70 m/s. Finally, we present a Beta regression outage modeling framework to address the shortcomings of existing power outage models.
... Many researchers (Liu et al., 2005(Liu et al., , 2007Han et al., 2009b;Guikema et al., 2010Guikema et al., , 2014Shashaani et al., 2018) have developed hurricane power outage prediction models to help utilities plan ahead of a storm for rapid deployments of resources and crews to expedite recovery from a storm. These models use input parameters, including hurricane winds, environmental parameters, power system information, and demographics to predict hurricaneinduced outages (Arora and Ceferino, 2022). ...
... 2), we need the total customer power interruption duration. Liu et al. (2007) modeled the time to recover the storm-induced power outages as a function of outage size since cities with more outages recover more slowly. Similarly, we model total customer power interruption duration as a function of outages. ...
... Many researchers (Liu et al., 2005(Liu et al., , 2007Han et al., 2009b;Guikema et al., 2010Guikema et al., , 2014Shashaani et al., 2018) have developed hurricane power outage prediction models to help utilities plan ahead of a storm for rapid deployments of resources and crews to expedite recovery from a storm. These models use input parameters, including hurricane winds, environmental parameters, power system information, and demographics to predict hurricaneinduced outages (Arora and Ceferino, 2022). ...
... 2), we need the total customer power interruption duration. Liu et al. (2007) modeled the time to recover the storm-induced power outages as a function of outage size since cities with more outages recover more slowly. Similarly, we model total customer power interruption duration as a function of outages. ...
Utility companies are often prepared for small-scale blackouts under normal operating conditions. However, they face crucial challenges with extreme weather events, such as hurricanes. Utilities must make risk-informed decisions to prioritize their limited resources (e.g., for grid hardening) in cities expected to experience larger and longer hurricane-induced outages. Probabilistic outage models can capture the uncertainty in power outages and characterize the grid's vulnerabilities to local environmental conditions, such as winds responsible for falling trees and poles that can affect the outages. We employ a probabilistic outage model developed for the 3.6 million historical outages caused by Hurricane Harvey (2017), Hurricane Michael (2018), and Hurricane Isaias (2020) in the United States states to investigate the expected performance of power systems to future hurricanes. This paper presents a generalized probabilistic framework coupling the expected frequency of future hurricane hazard levels and probabilistic outage predictions to understand the frequency of power system performance indices, namely the System Average Interruption Frequency Index and System Average Interruption Duration Index, from hurricanes. Results show that factors other winds, such as land use patterns, influence the risk of power outages across cities in New Jersey from future hurricanes.
... The most frequently used evaluation methods based on the quantification of system performance are quantitative methods, such as statistical analysis methods [16], Graph theory-based analysis methods [17], fuzzy logic model method [18], etc. Ref. [19] describes a specific formula for quantifying resilience widely used in studies. The proposed resilience curve is an indicator used to develop resilience, which is shown in Fig. 2. The operational performance of the power system is divided into four phases according to its characteristics as several changes occur in response to events. ...
... These data are mainly based on soil moisture, precipitation index, annual and monthly precipitation [29], etc. In [16], an accelerated failure time(AFT) model is proposed to help better inform customers and the public of expected post-storm power restoration times before damage assessments are performed in the field. Ref. [30] describes a Brownian motion model for load loss and outage interval prediction. ...
With the increasing attention to residential and industrial electricity reliability, resilience recovery of smart grids under extreme events, such as intentional attacks and natural calamities, has been considered a crucial technology. Power system, integrated with renewable energy, built complex and multielement energy structures to realize low carbon emission and clean. In recent years, there has been a lack of favorable research on resilience under renewable power systems, which is necessary to be sorted out in detail. Thus, this paper address this issue by exploring scientific consensus on principle, and diverse strategies of resilience under the circumstance of conventional and high proportions of renewable energy. It also presents the perspective of a hybrid equipment combination strategy to achieve power system recovery. Finally, this paper expounds on the prospects and opportunities in the context of power system resilience from the viewpoints of existing renewable energy devices and emerging technologies.
... It is important to recognize that various factors in each country affect the time required to restore power systems following an incident. These factors include geographical regions prone to natural phenomena [34], the state of existing infrastructure [35], and the management systems employed by different suppliers [36,37]. Consequently, proposing a universal solution for improving all energy supply systems in a cost-effective and replicable manner is impractical, given the significant variations in conditions across countries and energy service providers. ...
The studies on strategies for improving restoration times in electrical distribution systems are extensive. They have theoretically explored the application of mathematical models, the implementation of remotely controlled systems, and the use of digital simulators. This research aims to connect conceptual studies and the implementation of improvements and impact assessment in electrical distribution systems in developing countries, where distribution technologies vary widely, by employing a comprehensive methodology. The proposed research examines the restoration times for faults in substations within general distribution networks in the central-western region of Mexico. The study comprises these stages: (a) diagnosing the electrical supply, demand, and infrastructure; (b) analyzing the electrical restoration time and the restoration index of the substations; and (c) providing recommendations and implementing pilot tests for improvements in the identified critical substations. The results revealed 12 analysis zones, including 120 distribution substations, 150 power transformers, and 751 medium voltage circuits. Among the substations, 73% have ring connections, 15% have TAP connections, and 12% have radial connections. Additionally, 27% of the substations rely on only a single distribution line. The study identified areas with significant challenges in restoring electricity supply, particularly focusing on power transformers: 32 transformers with permanent power line failures requiring load transfer via medium voltage; 67 transformers requiring optimized restoration maneuvers due to specific characteristics; and 4 areas with opportunities to enhance the reliability of the power supply through remote-controlled link systems. The analysis resulted in the installation of 145 remote link systems, which improved restoration rates by over 40%. This approach is expected to be replicated throughout Mexico to identify improvements needed in the national distribution system.
... Barabadi et al. (2011) used the PRM model to analyze the influence of time-related covariates in mining on the recovery rate of crushing equipment. Liu et al. (2007) analyzed the difference between two common survival analysis models those are the accelerated failure time model and the Cox proportional hazard model. Showkat and Singh (2022) valuated the failure times of asphalt mixes in the indirect tensile strength tests using survival analysis based on Cox proportional hazards model. ...
Weather condition is an important factor affecting the air transportation, and it is essential for improving the air transportation capacity to identify the resilience recovery mechanism of civil airport infrastructure under weather extremes. In this paper, the time-varying model for resilience recovery of civil airport infrastructure under weather extremes is proposed based on the Cox proportional hazard model. The flight weather extremes events of the USA civil aviation flight data of 2019 are selected to establish the database for the analysis of resilience recovery of civil airport infrastructure, which includes 10 covariates. The statistical significance and risk rate of the covariates are investigated by the single factor method and PH assumption. The influence of covariates on the resilience level and recovery time of the civil airport infrastructure is explored. The results indicate that the rainfall, snow depth, delayed flight volume, low temperature, and crosswind are significant factors for the resilience recovery of civil airport infrastructure, and the regression coefficients of low temperature and crosswind are larger than the others. The key period for functional recovery of civil airport infrastructure is within 900 min after the weather extremes occurs. When the resilience function reaches to 50%, the recovery time of the infrastructure system increases by 30.0% and 26.2% considering the low temperature and crosswind respectively.
... Our evaluation encompassed various predictions, considering machine learning models such as random forest (RF), adaptive similar day (ASD), the combination of RF and ASD, support vector machines (SVM), and artificial neural networks (ANN), with different configurations relying on the evaluation metric of Accuracy over various datasets. Worthy of attention, there are various types of regression-based and statistical methods [14,34,45,54,57,[78][79][80][81] in the literature for power outage duration predictions with different evaluation metrics, such as MSE, MAE, MAPE, and RMSE, that are inappropriate to be compared with the current classification task. ...
In the face of increasing climate variability and the complexities of modern power grids, managing power outages in electric utilities has emerged as a critical challenge. This paper introduces a novel predictive model employing machine learning algorithms, including decision tree (DT), random forest (RF), k-nearest neighbors (KNN), and extreme gradient boosting (XGBoost). Leveraging historical sensors-based and non-sensors-based outage data from a Turkish electric utility company, the model demonstrates adaptability to diverse grid structures, considers meteorological and non-meteorological outage causes, and provides real-time feedback to customers to effectively address the problem of power outage duration. Using the XGBoost algorithm with the minimum redundancy maximum relevance (MRMR) feature selection attained 98.433% accuracy in predicting outage durations, better than the state-of-the-art methods showing 85.511% accuracy on average over various datasets, a 12.922% improvement. This paper contributes a practical solution to enhance outage management and customer communication, showcasing the potential of machine learning to transform electric utility responses and improve grid resilience and reliability.
... Their study highlighted the economic impacts of hurricane disruptions and the importance of pre-disaster mitigation planning. Liu et al. (2007) developed statistical models to predict power outage durations and electric grid restoration times following hurricanes and ice storms. Their study highlights the value of data-driven models to estimate recovery timelines for storm-disrupted grids. ...
The reliability of electric vehicle (EV) charging infrastructure during natural disasters is paramount for ensuring effective evacuation and response strategies, especially in regions vulnerable to such events. As EVs increasingly dominate the transportation landscape, disruptions to their essential infrastructure, such as charging stations, can profoundly hinder evacuation efficiency and safety. This research delves into the vulnerabilities of the EV charging network during a county-wide evacuation in Florida, a state regularly confronted by hurricanes and tropical storms. Using a multilayered network analysis, the study focuses on the electric grid robustness, shedding light on intricate interdependencies within the network. Central to this investigation are resiliency metrics that quantify the electric grid and charging stations’ functionality during and after natural disasters. Real-world data on EV ownership, supporting infrastructure, and historical disaster impacts further refine our realistic simulation scenarios. Our findings spotlight potential critical vulnerabilities across the electric infrastructure that may impact EV evacuation efforts during a natural disaster. These insights are poised to guide policymakers, urban planners, and emergency response teams in crafting strategies that fortify Florida’s EV charging infrastructure’s resilience during expansive evacuations.
... It is well-known that hurricanes and storms are one of the main reasons for wide-area and prolonged electrical outages. The operation of countless infrastructure systems can be affected adversely ranging from financial transactions to heating, security systems, water distribution, and business operations, and other services [78]. Violently high-speed winds and/or airborne debris may cause transmission lines to be downed and physical damage in distribution poles may occur [79]. ...
... Van et al. (2011) decoupled the problem into a multi-stage optimization problem by considering the time limits of some constraints. Chang et al. (1996) and Liu et al. (2007) predicted system performance response under other disasters by fitting recovery curves. NODA (1993), Sato and Ichii (1996), Xu et al. (2007), and Yan and Shih (2012) used an artificial intelligence approach to search for the optimal failure nodes recovery sequence. ...
The geomagnetically induced current (GIC) produced during extreme geomagnetic storms can easily lead to large‐scale blackouts in China due to the increase in the scale of its electric power grid. A power grid's resilience is its capability to resist various natural hazards, withstand primary failures, and quickly resume normal operation. To avoid power grid damages, this study developed a resilient power grid, incorporating failure, power flow calculation and recovery models under a uniform induced geoelectric field. We chose a system's performance loss as the resilience evaluation indicator, which intuitively reflected a system's loss under GIC. In addition, the recovery model was optimized using a genetic algorithm, and two resilience improvement measures were proposed. The IEEE‐RTS‐79 system, consisting of 10 generators, 24 buses and 5 transformers, was chosen as an example to verify the feasibility of this study. The results show that the genetic algorithm and optimization measures effectively enhanced the system's resilience indicator and provided a reference for preventing system damages under GIC and quick recovery after possible failures.
... Researchers have utilized data-driven approaches to predict the threats to power systems under EWE [8], [9]. The authors implemented cumulative time failure OPM utilizing the rainfall data of hurricane, maximum wind speed, and wind speed duration [10]. In [11], the authors implemented an approach for distribution networks to predict the risk level of power outage considering weather factors but without considering factors such as the region of study. ...
... The existing research on the weather-grid impact model can be categorized into statistical and simulation-based models. These models include a detailed modeling of power systems, extreme weather events, damage assessment, and restoration after extreme natural events [156][157][158]. ...
Over the past decade, extreme weather events have significantly increased worldwide, leading to widespread power outages and blackouts. As these threats continue to challenge power distribution systems, the importance of mitigating the impacts of extreme weather events has become paramount. Consequently, resilience has become crucial for designing and operating power distribution systems. This work comprehensively explores the current landscape of resilience evaluation and metrics within the power distribution system domain, reviewing existing methods and identifying key attributes that define effective resilience metrics. The challenges encountered during the formulation, development, and calculation of these metrics are also addressed. Additionally, this review acknowledges the intricate interdependencies between power distribution systems and critical infrastructures, including information and communication technology, transportation, water distribution, and natural gas networks. It is important to understand these interdependencies and their impact on power distribution system resilience. Moreover, this work provides an in-depth analysis of existing research on planning solutions to enhance distribution system resilience and support power distribution system operators and planners in developing effective mitigation strategies. These strategies are crucial for minimizing the adverse impacts of extreme weather events and fostering overall resilience within power distribution systems.
... Certainly, understanding the effect of weather extremes on the vulnerability of the electric grid is essential for quantifying trends of weather-related outages, and justifying investment to maintain and/or improve the reliability of the power system. Several studies of power system reliability under extreme weather conditions employed models and statistical methods to predict failure rates and restoration times (Duffey and Ha 2009;Liu et al. 2007;Zorn and Shamseldin 2015), transmission line outages (Dobson and Carreras 2010;Iešmantas and Alzbutas 2019;Xie et al. 2019), and applied a large body of machine learning (ML) models for power outage prediction (Arif and Wang 2018;Khodaei 2016, 2017;He and Cheng 2021;He, Cheng, Fang, and Crow, 2018;Yue et al. 2017). The foregoing efforts are based on recent weather events, which are encapsulated by the available data records from power utilities. ...
This study presents a framework for evaluating the vulnerability of the electrical grid to storm outages, based on multi-year atmospheric reanalysis datasets. The underlying methodology encompasses the classification of outage event severity and machine learning-based outage prediction models (OPMs), toward identifying relevant weather events and quantifying the associated outages within the European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5 and ERA5-Land) records. The proposed framework is tested for the Eversource Energy distribution grid over the State of Connecticut, for a period from 1981 to 2020. Within this context, and using as benchmark outage data reported by the utility for the period from 2005 to 2020, the accuracy of the classification for events of high-impact and extreme-severity proved to be high (i.e., 0.84 and 0.95, respectively). Especially for the latter case, the OPMs exhibited acceptable mean absolute percentage errors and high coefficient of determination (R²) values. Further, an analysis based on the annual maxima of the total number of outages, as well as the number of events with outages above various thresholds, indicated an intensification of extreme events over the last decade. Within this context, and given its importance to long-term planning and investment, we ultimately assess the potential impact of climate change on the resilience of the distribution grid, by evaluating the non-exceedance probabilities of six historical hurricanes that impacted the Eversource Energy service territory in Connecticut through a parametric statistical approach.
... In recent years, with more and more data collected from Electric power companies, datadriven methods have been applied to power system fault prediction under typhoon disasters. Reference [21] applied the accelerated failure time model to predict the downtime in the case of hurricane and ice storm, and the model performance is compared with different combinations of input variables. Guikema et al. employed a data mining method to predict the number of customers without power based on meteorological data, geographic data and power grid data [22]. ...
Heavy rains pose a great threat to the reliable and secure power supply of urban distribution networks. A knowledge-informed data-driven resilience assessment approach is proposed to evaluate urban distribution networks’ abilities to resist heavy rains. Firstly, the rainstorm waterlogging process is simulated to obtain the rainstorm intensity and rainfall process, providing the input for the data-driven model. Then, input variables are grouped guided by expert knowledge, and a dynamic and static data-driven model is constructed to predict the line outages based on historical data. Finally, the Monte Carlo sismulation method integrated with the data-driven model is developed to assess the resilience of urban distribution networks and the number of line outages is selected as the evaluation metric. The effectiveness of the proposed method is sufficiently validated by the historical data of an urban distribution network.
... Models and methods developed by Han et al. (2009), Nateghi et al. (2011 and Liu et al. (2007), for example, utilize data from numerous historical hurricanes and ice storms to predict outage times and locations for future storms on the Gulf and East Coasts of the United States. These estimations can be employed as a storm approaches to help residents and operators prepare before the power system is disrupted. ...
Earthquakes and other natural hazards can cause significant damage to structures and critical lifeline systems. To effectively prepare for disasters, planners and emergency managers can use modeling to assess the impacts that a disaster could have at a municipal and regional level. Modeling approaches are often technically complex and require a great deal of data and expertise to develop and assess. Technical assessments are immensely valuable for providing a detailed understanding of a system, but their complexity makes it challenging to provide opportunities for engagement with a variety of audiences and to compare different scenarios and their effects on a region. The work presented here seeks to demonstrate multiple assessment methods and their utility for planning purposes by using a model that tracks infrastructure system dependencies, repair times, and resource requirements. The assessment methods include developing recovery curves that can be used to assess outage effects on communities, ranking recovery times for different zones to describe areas that are relatively more or less at risk after a disaster, and comparing system recovery time to repair time to assess internal and external dependencies. This work provides an overview of the modeling approach and its representation of water, wastewater, power, and road and highway systems. A case study of a simulated earthquake and its effect on the Metro Vancouver region of British Columbia, Canada, is also presented and examples of the utility of each assessment methodology are detailed. The goal of this work is to provide additional resources for planners and policy makers so that they are equipped to make decisions that best protect their communities.
... In Mensah and Dueñas-Osorio (2014), wind fragility curves are incorporated in a Bayesian network model driven by PF modeling to predict outages caused by hurricane winds. Data-driven predictive approaches such as accelerated failure time models (Liu et al., 2007) and random forest models (Guikema et al., 2014) have also been used to assess wind-oriented hurricane consequences. ...
We present a stochastic programming model for informing the deployment of temporary flood mitigation measures to protect electrical substations prior to an imminent and uncertain hurricane. The first stage captures the deployment of a fixed number of mitigation resources, and the second stage captures grid operation during a contingency. The primary objective is to minimize expected load shed. We develop methods for simulating flooding induced by extreme rainfall and construct two geographically realistic case studies, one based on Tropical Storm Imelda and the other on Hurricane Harvey. Applying our model to those case studies, we investigate the effect of the mitigation budget on the optimal objective value and solutions. Our results highlight the sensitivity of the optimal mitigation to the budget, a consequence of those decisions being discrete. We additionally assess the value of having better mitigation options and the spatial features of the optimal mitigation.
... Outage restoration in natural disasters differs from individual outage restoration on a clear day because outages can be widespread and compounding [6], which lead to difficult restoration scenarios that require expert management of available resources. Available outage restoration models exist with techniques such as expert systems [7], fuzzy logic [8], heuristic approaches [9], [10], generic restoration milestones [11], Monte Carlo simulation [12], accelerated failure time [13], deterministic methods [14], [15], and optimization models [16]. These existing models lack scenario modeling and are limited in their usability as a predictive tool. ...
p>The increasing frequency and intensity of high impact storms, especially in Northeast United States, requires utilities and emergency managers to be increasingly prepared for lengthy power outage restorations. Historically, restoration has relied on emergency managers decennial experience with limited access to predictive models. This study highlights the development of a combined system composed of the UConn Outage Prediction Model (OPM) for predicting weather-related damage in the distribution system and an Agent-Based Model (ABM) for estimating the time to electric power restoration. The combined system is validated using Outage Management System (OMS) and crew deployment information for four historical extreme weather events that occurred in the State of Connecticut in the past decade. Through the ABM’s ability to test different restoration strategies, we study the impact that human knowledge and decisions have on the outage restoration curve. Furthermore, we use the model to test how the restoration could have been different if crews were allocated to area work centers based on the location of damage predictions from the UConn OPM and on increased crew counts, reflecting a more aggressive storm preparedness. This test highlights how an OPM-ABM system can benefit emergency preparedness and response managers in advance of storms impact. </p
... Other statistical approaches involved predicting the distribution of outage duration after major storms using accelerated failure models within the GAM framework (Liu, Davidson & Apanasovich 2007). This approach could also be used to predict individualised times to overhead distribution lines' failures, thus allowing for targeted maintenance scheduling. ...
Overhead distribution lines play a vital role in distributing electricity, however, their freestanding nature makes them vulnerable to extreme weather conditions and resultant disruption of supply. The current UK regulation of power networks means preemptive mitigation of disruptions avoids financial penalties for distribution companies, making accurate fault predictions of direct financial importance. Here we present predictive models developed for a UK network based on gradient-boosted location, scale, and shape models, providing spatio-temporal predictions of faults based on forecast weather conditions. The models presented are based on (a) tree base learners or (b) penalised smooth and linear base learners -- leading to a Generalised Additive Model (GAM) structure, with the latter category of models providing best performance in terms of out-of-sample log-likelihood. The models are fitted to fifteen years of fault and weather data and are shown to provide good accuracy over multi-day forecast windows, giving tangible support to power restoration.
... Estimation of restoration time for electric power systemsLiu et al. (2007) General natural and man-made hazardsRestoration program for infrastructure networks Hu et al. (2016), Safapour et al. (2020) Restoration of critical infrastructure Fang and Sansavini (2019), Vugrin et al. (2010) Restoration of transportation networks Chen and Miller-Hooks (2012), Fang and Sansavini (2017), Liao et al. (2018), Vugrin et al. (2014) Optimal scheduling of emergency roadway repair Hayat et al. (2019), Yan and Shih (2009) Restoration of bridges along highways Bocchini and Frangopol (2012) Ranking of repair schedules for water distribution system Bałut et al. (2019) Restoration under uncertainty for power grids Fang and Sansavini(2019) ...
... Perhaps because of these technical difficulties, previous research in outage modeling and forecasting has often neglected to specifically address the technical problems with predicting extreme events, despite the popularity of research that focuses specifically on tropical storm related outage modeling. Much of the diversity of this literature is in the modeling algorithms, ranging from linear models (Han et al., 2009b,a;Liu et al., 2005;Quiring et al., 2014), Bayesian analysis (Yue et al., 2018;Alpay et al., 2020), to other machine learning approaches (Guikema et al., 2010;Liu et al., 2007;McRoberts et al., 2018;Guikema and Quiring, 2012;Nateghi et al., 2011Nateghi et al., , 2014aQuiring et al., 2014Quiring et al., , 2011Wanik et al., 2018;Kankanala et al., 2014;Alpay et al., 2020). ...
Power outages caused by extreme weather events cost the economy of the United States billions of dollars every year and endanger the lives of the people affected by them. These types of events could be better managed if accurate predictions of storm impacts were available. While empirical power outage prediction models have been in development for many years, accurate operational predictions of the most extreme and impactful weather-related outage events have proven difficult to achieve for several reasons. In this paper, we describe a data intensive modeling approach specifically designed for forecasting the impacts of extreme weather events on power distribution grids. To that end, methods for developing datasets that include a large number of example storms and predictors are described. In addition, we test several methods of managing the extreme value distribution of the target variable via statistical transformation and balancing of the dataset. The best performing outage prediction model developed here is capable of predicting storm impacts across four orders of magnitude with R2 and Nash-Suttcliffe Efficiency scores of 0.82. Also, by investigating the model’s sensitivities and predictions for the highest impact events, we find that there is significant diversity in the meteorological factors that drive the predictions for the most severe events, suggesting that the weather hazards are more complex than they often treated in empirical analyses of their impacts. The accuracy of the outage model, together with the importance of various meteorological variables that contribute to that accuracy, validate the described methodology and suggest that future empirical analysis of the impacts of extreme weather should include multifaceted descriptions of the hazard to better represent the complex factors which contribute to the most impactful events.
... Outage restoration in natural disasters differs from individual outage restoration on a clear day because outages can be widespread and compounding [6], which lead to difficult restoration scenarios that require expert management of available resources. Available outage restoration models exist with techniques such as expert systems [7], fuzzy logic [8], heuristic approaches [9], [10], generic restoration milestones [11], Monte Carlo simulation [12], accelerated failure time [13], deterministic methods [14], [15], and optimization models [16]. These existing models lack scenario modeling and are limited in their usability as a predictive tool. ...
The increasing frequency and intensity of high impact storms, especially in Northeast United States, requires utilities and emergency managers to be increasingly prepared for lengthy power outage restorations. Historically, restoration has relied on emergency managers decennial experience with limited access to predictive models. This study highlights the development of a combined system composed of the UConn Outage Prediction Model (OPM) for predicting weather-related damage in the distribution system and an Agent-Based Model (ABM) for estimating the time to electric power restoration. The combined system is validated using Outage Management System (OMS) and crew deployment information for four historical extreme weather events that occurred in the State of Connecticut in the past decade. Through the ABM’s ability to test different restoration strategies, we study the impact that human knowledge and decisions have on the outage restoration curve. Furthermore, we use the model to test how the restoration could have been different if crews were allocated to area work centers based on the location of damage predictions from the UConn OPM and on increased crew counts, reflecting a more aggressive storm preparedness. This test highlights how an OPM-ABM system can benefit emergency preparedness and response managers in advance of storms impact.
... In [8], a theoretical tutorial system is proposed to train distribution system operators to effectively respond to emergencies. The study in [9] proposes a cooperative agents-based system for service restoration through artificial intelligence methods. In [10], a framework using a fuzzy logic is developed to manage outages. ...
Microgrid formation is a promising solution to enhance resiliency of distribution networks. The self-adequacy feature of a microgrid enables continuity of power supply through distributed generation (DG) units during severe faults and natural disasters. In this paper, different methods commonly used to partition a distribution network into multiple microgrids are presented, including the graph theory, heuristic rule-based algorithm, cluster-based technique, and mixed integer programming. Advantages and disadvantages of these techniques and future research directions are presented. This review provides an excellent summary on service restoration through micrgrid formation, and offers a valuable reference for researchers working on grid modernization of distribution networks.
... In order to have a clear comparison between the pre-event and post-event resilience levels of a power system, a quantitative assessment is preferred, and more attention is paid to it in the related literature. Quantitative methods for resilience assessment can be mainly classified into statistical methods [53], risk-based methods [35], system fragility-based methods [62], graph theory-based methods [14], simulation-based methods, and fuzzy logic models [39] and they involve both operational and infrastructure mathematical formulations of resilience metrics. Generally, the model presented in [52] is used for a quantitative evaluation of resilience. ...
Les systèmes critiques tels que les systèmes de calcul distribués nécessitent une grande fiabilité et résilience pour garantir la qualité du service. De plus, 90% de toutes les interruptions de service subies par les utilisateurs proviennent du système d’alimentation et de distribution électrique. En effet, les systèmes de distribution existants sont centralisés et reposent principalement sur le réseau public. D’autre part, la technologie des microgrid DC est de plus en plus répandue dans les bâtiments, les navires et les centres de données. Il promet un changement d’orientation de la génération centralisée vers une co-génération distribuée et plus verte. Cependant, les architectures de microgrid DC existantes sont statiques, ce qui signifie qu’elles ne peuvent pas changer leur topologie après l’installation. Cet aspect limite la flexibilité et l’adaptabilité du réseau électrique dans un tel scénario hétérogène et variable.Cette thèse propose une nouvelle architecture de microgrid DC qui permet le concept de "Software-Defined Power Domains". En effet, en utilisant le crossbar de puissance conçue, la topologie peut être modifiée dynamiquement par logiciel. Cet aspect améliore la flexibilité du système de distribution d’électricité, qui peut se reconfigurer pour mieux s’adapter à l’état des charges et des sources. En outre, la disponibilité du système est largement améliorée en raison de la redondance du bus et de la segmentation fournie par le crossbar. Enfin, la résilience augmente grâce aux opérations dynamiques capables d’isoler instantanément une défaillance, de reconfigurer la topologie et de restaurer le fonctionnement du système.Dans cette thèse, tous les détails de cette architecture sont fournis. Ensuite, le contrôle et certaines opérations dynamiques sont explorés et testés à l’aide du simulateur PSIM pour valider les hypothèses susmentionnées. De plus, la fiabilité et la disponibilité du système sont largement analysées à l’aide des chaînes de Markov ou des méthodes Monte Carlo lorsqu’on considère des composants avec des taux de défaillance constants ou non constants, respectivement. Enfin, le cas d’étude d'un centre de données alimenté en courant continu sert à valider les avantages de l’architecture proposée. Plusieurs transformations sont appliquées à l’architecture de base qui améliore progressivement la disponibilité et le MTBF du système. Par exemple, l’architecture basée sur crossbars en configuration matricielle augmente le MTBF du système de 4 fois.
... The weather-related power failure of an entire grid must be predicted to accommodate increased size and interconnectivity of an electric power grid. The power failure prediction can be achieved mainly by models based on regressive statistical methods (Liu 2015;Liu, Davidson, and Apanasovich 2007;Aznarte and Siebert 2016). Once an appropriate weather model is determined, a suitable index is required to measure the resilience of the system. ...
The frequency of extreme weather events and the resulting impact on the electric grid has grown in recent years. This paper reviews previous work on distribution system resilience, concentrating on electrical network protection in the face of extreme events. Firstly, it analyses confounding terminology used in power systems resilience studies, such as definitions, resilience vs reliability, and resilience curve. Secondly, resilience techniques are examined to understand better the influence of extreme events on the probability and duration of electrical network failures. The difficulties involved in the resilience quantification and analysis process due to several perceptions and complexity in the characterisation of extreme events are pointed out. Thirdly, this research divides current resilience enhancement techniques into two categories: planning-based and operational-based resilience. It contributes to a comprehensive study of various methods in each of those categories. Finally, the research gaps and possible solutions to existing distribution system resilience methods are provided.
Hurricanes can cause devastating damage to overhead distribution lines leading to large power outages in electric grids. Power outage prediction models can help utilities to plan for an expedited power recovery by identifying the extent of power disruptions before the arrival of a hurricane. These models often use multiple input parameters, including early warning forecasts of hurricane characteristics, environmental data, power system details, and demographic information. We propose a quasi-binomial regression model to advance power outage models and overcome their existing limitations, such as unbounded outage predictions, limited extrapolation, and high uncertainties at low and high winds. This paper shows that the quasi-binomial model allows us to better capture the mechanics of power system failures due to hurricanes. We fitted our model to power outage data across 2,322 cities for four historical hurricanes: Harvey (2017), Michael (2018), Isaias (2020), and Ida (2021). We validated our model for the outages in Florida during Hurricane Ian (2022). The quasi-binomial model outperformed existing random forest and negative binomial regression models with a 7% error versus 50% and 76%, respectively. To demonstrate the quasi-binomial model’s good performance more comprehensively, we also tested a new beta regression model for outages. We show the quasi-binomial model had a smaller cross-validation root-mean squared error of 0.23 compared with 0.28 for the beta model. Finally, we show that our model also captures that grids with more redundant components can be more resilient to hurricane-caused outages. Thus, our proposed quasi-binomial model advances the state of the art for power outage predictions.
A wide variety of weather conditions, from windstorms to prolonged heat events, can substantially impact power systems, posing many risks and inconveniences due to power outages. Accurately estimating the probability distribution of the number of customers without power using data about the power utility system and environmental and weather conditions can help utilities restore power more quickly and efficiently. However, the critical shortcoming of current models lies in the difficulties of handling (i) data streams and (ii) model uncertainty due to combining data from various weather events. Accordingly, this article proposes an adaptive ensemble learning algorithm for data streams, which deploys a feature‐ and performance‐based weighting mechanism to adaptively combine outputs from multiple competitive base learners. As a proof of concept, we use a large, real data set of daily customer interruptions to develop the first adaptive all‐weather outage prediction model using data streams. We benchmark several approaches to demonstrate the advantage of our approach in offering more accurate probabilistic predictions. The results show that the proposed algorithm reduces the probabilistic predictions' error of the base learners between 4% and 22% with an average of 8%, which also result in substantially more accurate point predictions. The improvement made by our algorithm is enhanced as we exchange base learners with simpler models.
Strong hurricane winds damage power grids and cause cascading power failures. Statistical and machine learning models have been proposed to predict the extent of power disruptions due to hurricanes. Existing outage models use inputs including power system information, environmental parameters, and demographic parameters. This paper reviews the existing power outage models, highlighting their strengths and limitations. Existing models were developed and validated with data from a few utility companies and regions, limiting the extent of their applicability across geographies and hurricane events. Instead, we train and validate these existing outage models using power outages from multiple regions and hurricanes, including hurricanes Harvey (2017), Michael (2018), and Isaias (2020), in 1910 US cities. The dataset includes outages from 39 utility companies in Texas, 5 in Florida, 5 in New Jersey, and 11 in New York. We discuss the limited ability of state-of-the-art machine learning models to (1) make bounded outage predictions, (2) extrapolate predictions to high winds, and (3) account for physics-informed outage uncertainties at low and high winds. For example, we observe that existing models can predict outages higher than the number of customers (in 19.8 % of cities with an average overprediction ratio of 5.2) and cannot capture well the outage variance for high winds, especially above 70 m s-1. Our findings suggest that further developments are needed for power outage models for proper representation of hurricane-induced outages.
This paper develops a data-driven approach to accurately predict the restoration time of outages under different scales and factors. To achieve the goal, the proposed method consists of three stages. First, given the unprecedented amount of data collected by utilities, a sparse dictionary-based ensemble spectral clustering (SDESC) method is proposed to decompose historical outage datasets, which enjoys good computational efficiency and scalability. Specifically, each outage sample is represented by a linear combination of a small number of selected dictionary samples using a density-based method. Then, the dictionary-based representation is utilized to perform the spectral analysis to group the data samples with similar features into the same subsets. In the second stage, a knowledge-transfer-added restoration time prediction model is trained for each subset by combining weather information and outage-related features. The transfer learning technology is introduced to deal with the underestimation problem caused by data imbalance in different subsets, thus improving the model performance. Furthermore, to connect unseen outages with the learned outage subsets, a t-distributed stochastic neighbor embedding-based strategy is applied. The proposed method fully builds on and is also tested on a large real-world outage dataset from a utility provider with a time span of six consecutive years. The numerical results validate that our method has high prediction accuracy while showing good stability against real-world data limitations.
The building group is the basis for the maintenance and operation of the city. The rapid recovery of community building group (CBG) can effectively reduce economic losses caused by earthquakes. There is service function interdependence among the buildings, and the impact of this interdependence on the postdisaster recovery of CBG is not clear. In order to improve the postdisaster recovery efficiency of CBG and explore the impact of service function dependence among buildings on postdisaster recovery, this paper integrates the recovery time (RT) and functionality loss (FL) of individual buildings and proposes two multi‐objective optimization recovery models suitable for CBG: the postearthquake recovery optimization model of CBG and the postearthquake recovery optimization model of CBG considering the service function interdependence among buildings. In these models, building RT, building FL, CBG resilience, and building recovery resources are considered, and the nondominated sorting genetic algorithm‐Ⅱ is used to solve the models to obtain the optimal restoration scheduling of CBG. At the same time, through a case, this paper analyzes the impact of the interdependence of service function among buildings and recovery resources on the recovery scheduling and resilience of CBG. This method can provide a basis for pre‐earthquake disaster risk reduction planning and significantly improves the postdisaster recovery efficiency of CBG.
The prediction of typhoon-induced transmission line outages is essential to improve the resilience of urban power systems. This paper proposes a novel data-driven prediction model to promote the accuracy by quantifying the cumulative influence of dynamic data and mitigating the data imbalance. In the model, the static data and the dynamic data compose the disaster-causing feature vector as model input. Then, the denoising adaptive synthetic (ADASYN) sampling algorithm is proposed to select target samples purposely and generate minority samples adaptively to balance the dataset. Also, the discriminative model guarantees the consistency of the data distribution. Thereby, the dual path model is proposed to quantify the stable impact of static data and cumulative impact of dynamic data based on the feedforward neural network and gated recurrent unit (GRU), respectively, and fuse the extracted features with the multi-head attention mechanism to predict the category of the number of line outages. Based on the real dataset, this paper compares the performance of the denoising ADASYN algorithm and dual path model with benchmarking algorithms. The experiment results indicate that the proposed method witnesses a better accuracy in predicting typhoon-induced transmission line outages.
This paper presents a data-driven tower damage prediction model to predict the damage spatial arrangement of 10 kV towers under typhoon disaster. The 10 kV tower belongs to distribution network in China. Compared with high voltage level in transmission network, the 10 kV power towers are more vulnerable to typhoon disasters due to their lower design criterion and large amount. The data-driven model proposed in this paper can effectively predict the tower damage situation. It takes meteorological, power grid and geographic information into account and can be divided into two steps. The first step is to predict the damage probability of each tower by using data-driven method such as AdaBoost, Gradient Boosting Regression, K Nearest Neighbor Regressor, Random Forest and Support Vector Regression algorithms. In this step, the data space is constructed by data processing and variable correlation analysis. Then, we use the processed meteorological, power grid and geographic information as input and the damage probability as output for model comparison. The second step is to select the optimal model based on comprehensive index weighting. Through a comprehensive comparison of the efficiency and accuracy of the five models in various actual scenarios, the optimal model is Gradient Boosting Regression, which outperforms the other adverse algorithms and produces the prediction damage consistent with actual data.
© 2017 Elsevier Inc. All rights reserved.
This thesis addresses several challenges corresponding to the operation and planning of electricity grids due to severe weather conditions. It considers electricity grid challenges due to natural vulnerabilities including wind speed, wind gusts, and wildfires. A reformulation for the principles of automatic generation control (AGC) in a decomposed convex relaxation algorithm is presented. It finds an optimal solution to the AC optimal power flow (ACOPF) problem that is secure against a large set of contingencies. A surrogate model to quantify the wildfire ignition by each power line under extreme weather conditions is presented. The system operator can use this model to de-energize
power lines during Public Safety Power Shutoffs (PSPS). A two-stage robust optimization
problem is formulated to ensure the risk-averse resilient operation of the electricity grid
under the risk of wildfire for 24 hours. The uncertainty in demand and solar generation
are incorporated into the formulation. A 10-year expansion planning of power system under fire hazard weather conditions to find improved balance for proactive de-energization of transmission lines, distributed solar integration, modification of transmission lines, and addition of new lines is presented. The validity of the presented surrogate model and the optimization problems are demonstrated in various test cases.
The increased complexity of infrastructure systems has resulted in critical interdependencies between multiple networks—communication systems require electricity, while the normal functioning of the power grid relies on communication systems. These interdependencies have inspired an extensive literature on coupled multilayer networks, assuming a hard interdependence, where a component failure in one network causes failures in the other network, resulting in a cascade of failures across multiple systems. While empirical evidence of such hard failures is limited, the repair and recovery of a network requires resources typically supplied by other networks, resulting in documented interdependencies induced by the recovery process. In this work, we explore recovery coupling, capturing the dependence of the recovery of one system on the instantaneous functional state of another system. If the support networks are not functional, recovery will be slowed. Here we collected data on the recovery time of millions of power grid failures, finding evidence of universal nonlinear behavior in recovery following large perturbations. We develop a theoretical framework to address recovery coupling, predicting quantitative signatures different from the multilayer cascading failures. We then rely on controlled natural experiments to separate the role of recovery coupling from other effects like resource limitations, offering direct evidence of how recovery coupling affects a system’s functionality. Infrastructure and power systems are often represented as multilayer structures of interdependent networks. Danziger and Barabási demonstrate the presence of recovery coupling in such systems, where the recovery of an element in one network requires resources from nodes and links in another network.
Hurricanes can cause extensive power outages, resulting in economic loss, business interruption, and secondary effects to other infrastructure systems. Currently, power companies are unable to accurately predict where outages will occur. Therefore, it is difficult for them to deploy repair personnel and materials, and make other emergency response decisions in advance of an event. This paper describes negative binomial regression models for the number of hurricane-related outages likely to occur in each one square kilometer grid cell and in each zip code in a region due to passage of a hurricane. The models are based on a large Geographic Information System database of outages in North and South Carolina from three hurricanes: Floyd (1999), Bonnie (1998), and Fran (1996). The most useful explanatory variables are the number of transformers in the area, the company affected, maximum gust wind speed, and a hurricane effect. Wind speeds were estimated using a calibrated hurricane wind speed model. Pseudo R-squared values and other diagnostic statistics are developed to facilitate model selection with generalized negative binomial models. Journal of Infrastructure Systems
There are many models for hindcasting ice loads from meteorological data measured during freezing rain storms. Each model is based on the physics of the ice accretion process and on empirical observations. However, these models predict significantly different ice loads for the same freezing rain storm, making it difficult to use model results to determine design ice loads. In this paper, we describe a simple ice load model that can be used to make conservative back-of-the-envelope calculations of ice loads based on the precipitation rate and wind speed. Using historical weather data from Springfield, IL, we compare the ice loads from this model with those from other models and discuss the reasons for the differences between them. We also compare the modeled and measured ice loads from one well-documented storm that occurred at CRREL's freezing rain weather station.
To have a national methodology for pre-earthquake planning, a model for predicting the post-earthquake behavior of city lifeline systems was developed. We discussed three factors in the model: structural damage, functional damage, and the restoration process after the earthquake. The restoration process is basically described by a differential equation applicable to a service area represented by a census mesh, and is applied to the lifelines (i.e., supply systems of gas, electric power, and water) of Sendai city in Japan. The model, in addition, indicates the lifeline network properties, and serviceability indices are defined in order to assess the functional damage of each system. In the case of the 1978 Off-myagi earthquake, a computer simulation of the restoration process was carried out by using step by step calculations and the Monte Carlo method. The simulated results, using indices as a function of time, were well in agreement with actual results, which indicates that the model is capable of predicting the restoration process. Through further simulations which varied the restoration strategies of the Emergency Headquarters, we show that the recovery of the gas system is sensitively affected by the strategy used. However, the electric power and the water systems were more influenced by the network properties rather than the strategies used. Our approach can provide useful information in undertaking pre-earthquake countermeasures for city lifeline systems.
This paper describes the application of a new discrete-event-simulation model of the post-earthquake electric power restoration process in Los Angeles. The findings are that (1) Los Angeles residents may experience power outages lasting up to 10 days; (2) what we call the power rapidity risk (the joint probability distribution of restoration of a specified number of customers in a specified amount of time) varies throughout the area; (3) there is a relatively high likelihood that more repair materials than are currently available will be required if a large earthquake occurs; and (4) there are ways to reduce the expected duration of earthquake-initiated power outages and they should be subjected to cost-benefit analysis. These results should be useful to utilities and emergency planners in Los Angeles. The new simulation modeling approach could be used in other seismically active cities to gain insights into the restoration process that other modeling approaches cannot provide.
The functioning of lifeline systems after a major earthquake is critical to the modern urban center. A system study of lifeline response to catastrophic earthquakes is the essential prerequisite for the emergency management authority to form a mitigation and reconstruction plan to minimize the total loss caused by the earthquake. This paper develops an applied formulation of lifeline-restoration processes in the post-earthquake period, on the basis of a discrete-state, discretetime Markov process. There are a finite number of damaged lifelines that each require a resource for reconstruction. The lifelines have values associated with their functioning capacity. We wish to assign limited resources to lifelines so as to minimize the total loss caused by malfunction of damaged lifelines. Dynamic programming is used to optimize the distribution of limited reconstruction resources. By computer simulation, various scenarios are examined, and useful information that is important to the emergency management authority is obtained.
Urban centers in the United States are more likely to experience lifeline disruption and failure following natural disasters than incurring massive loss of life or structural collapse (Okada (1997)). In this paper, the performance of an urban utility distribution system is analyzed for the two common hazards of earthquakes and extreme winds. The empirical results show how disruption conditions and restoration efforts are different for the two hazards. Consideration of the ductility of the lifeline network may identify strategies for improvement.
The Weather Research and Forecast (WRF) project is a multi-institutional effort to develop an advanced mesoscale forecast and data assimilation system that is accurate, efficient, and scalable across a range of scales and over a host of computer platforms. The first release, WRF 1.0, was November 30, 2000, with operational deployment targeted for the 2004-05 time frame. This paper provides an overview of the project and current status of the WRF development effort in the areas of numerics and physics, software and data architecture, and single-source parallelism and performance portability.
This monograph examines the potential effects of a repeat of the New Madrid earthquake to the metropolitan Memphis area. The authors developed a case study of the impact of such an event to the electric power system, and analyzed how this disruption would affect society. In nine chapters and 189 pages, the book traces the impacts of catastrophic earthquakes through a curtailment of utility lifeline services to its host regional economy and beyond. the monographs` chapters include: Modeling the Memphis economy; seismic performance of electric power systems; spatial analysis techniques for linking physical damage to economic functions; earthquake vulnerability and emergency preparedness among businesses; direct economic impacts; regional economic impacts; socioeconomic and interregional impacts; lifeline risk reduction; and public policy formulation and implementation.
THE MOST PRACTICAL, UP-TO-DATE GUIDE TO MODELLING AND ANALYZING TIME-TO-EVENT DATANOW IN A VALUABLE NEW EDITION Since publication of the first edition nearly a decade ago, analyses using time-to-event methods have increase considerably in all areas of scientific inquiry mainly as a result of model-building methods available in modern statistical software packages. However, there has been minimal coverage in the available literature to9 guide researchers, practitioners, and students who wish to apply these methods to health-related areas of study. Applied Survival Analysis, Second Edition provides a comprehensive and up-to-date introduction to regression modeling for time-to-event data in medical, epidemiological, biostatistical, and other health-related research. This book places a unique emphasis on the practical and contemporary applications of regression modeling rather than the mathematical theory. It offers a clear and accessible presentation of modern modeling techniques supplemented with real-world examples and case studies. Key topics covered include: variable selection, identification of the scale of continuous covariates, the role of interactions in the model, assessment of fit and model assumptions, regression diagnostics, recurrent event models, frailty models, additive models, competing risk models, and missing data. Features of the Second Edition include: Expanded coverage of interactions and the covariate-adjusted survival functions The use of the Worchester Heart Attack Study as the main modeling data set for illustrating discussed concepts and techniques New discussion of variable selection with multivariable fractional polynomials Further exploration of time-varying covariates, complex with examples Additional treatment of the exponential, Weibull, and log-logistic parametric regression models Increased emphasis on interpreting and using results as well as utilizing multiple imputation methods to analyze data with missing values New examples and exercises at the end of each chapter Analyses throughout the text are performed using Stata Version 9, and an accompanying FTP site contains the data sets used in the book. Applied Survival Analysis, Second Edition is an ideal book for graduate-level courses in biostatistics, statistics, and epidemiologic methods. It also serves as a valuable reference for practitioners and researchers in any health-related field or for professionals in insurance and government.
This paper presents a stochastic integer program developed to determine how to schedule inspection, damage assessment, and repair tasks so as to optimize the post-earthquake restoration of the electric power system. The objective of the optimization is to minimize the average time each customer is without power, and a genetic algorithm is used to solve it. The effectiveness of the schedules recommended by the optimization are evaluated by running a detailed discrete event simulation model of the restoration process with both the optimization-generated schedules and the power company's original schedules, and comparing the resulting restorations according to three measures—average time each customer is without power, time required to restore 90 of customers, and time required to restore 98 of customers. The optimization and simulation models both consider all the earthquakes that could affect the power system and represent the uncertainty surrounding expected restoration times. he models were developed through an application to the Los Angeles Department of Water and Power (LADWP) electric power system, but the general approach is extendable to other electric power systems, other lifelines, and other hazards. Copyright © 2006 John Wiley & Sons, Ltd.
This paper describes the development of event-based hurricane simulation techniques for the evaluation of long-term risks in the Southeastern United States. Numerical simulation techniques were used to evaluate the 50-yr mean recurrence interval gradient wind-speeds for hurricane-prone regions in the study area. Using a damage model derived from actual insurance loss data and developed by the authors, the expected annual insurance losses were evaluated. The state of South Carolina was used as a case study to demonstrate the applicability of this system.
This article describes flexible statistical methods that may be used to identify and characterize nonlinear regression effects. These methods are called "generalized additive models". For example, a commonly used statistical model in medical research is the logistic regression model for binary data. Here we relate the mean of the binary response ¯ = P (y = 1) to the predictors via a linear regression model and the logit link function: log
Thesis (Ph.D.)--Cornell University, August, 2006. Includes bibliographical references.
The 2004 Great Sumatra-Andaman earthquake had an average source duration of about 500 sec. and a rupture length of 1,200–1,300 km. The seismic moment, M0, determined with a finite source model, was 6.5×1022 N-m, which corresponds to Mw=9.18. Allowing for the uncertainties in the current M0 determinations, Mw is in the range of 9.1 to 9.3. The tsunami magnitude Mt is 9.1, suggesting that the overall size of the tsunami is consistent with what is expected of an earthquake with Mw=9.1 to 9.3. The short-period body-wave magnitude m-hatb is 7.25, which is considerably smaller than that of large earthquakes with a comparable Mw. The m-hatb versus Mw relationship indicates that, overall, the Great Sumatra-Andaman earthquake is not a tsunami earthquake. The tectonic environment of the rupture zone of the Great Sumatra-Andaman earthquake is very different from that of other great earthquakes, such as the 1960 Chile and the 1964 Alaska earthquakes. This difference may be responsible for the unique source characteristics of this earthquake. The extremely large size of the Great Sumatra-Andaman earthquake is reflected in the large amplitude of the long-period phase, the W phase, even in the early part of the seismograms before the arrival of the S wave. This information could be used for various early warning purposes.
The Pacific Gas and Electric Company (PGE), together with
California Polytechnic State University in San Luis Obispo, has
developed a personal computer program for evaluating the reliability
levels of their electric distribution circuits. The program, called the
Distribution Reliability Assessment Model (DREAM), incorporates
historical outage information with circuit component failure rates and
estimates of fault response and component repair times to compute
expected levels of customer outage frequency and duration. The authors
highlight recent enhancements to DREAM which allow for the evaluation of
both overhead and underground circuits. The authors also describe their
efforts to measure the accuracy of the DREAM calculations and the
application of the program in predicting the reliabilities of 180
feeders
Lightning is a significant cause of faults and outages in many electric power systems and is one of the major causes of poor system reliability. Predictive assessment of distribution reliability indices can be used to identify areas that have poor reliability so that appropriate changes in system design can be implemented. The assessment of distribution system performance under lightning conditions requires modeling of storm characteristics and system response. In this paper, a Monte Carlo simulation for evaluating distribution system reliability under lightning storm conditions is presented. The results from a practical distribution system show the importance of detailed modeling of storm characteristics and simulation of the system response in assessing distribution system reliability during lightning storms.
The goal of distribution system reliability assessment is to
predict the availability of power at each customer's service entrance.
Existing methods predict the interruption frequency and duration each
customer can expect, but omit two major contributing factors: momentary
interruptions and storms. This paper presents methods to determine the
impact of each phenomena. These methods are then used to assess the
reliability of an existing utility distribution system and to explore
the reliability impact of distribution automation
Performance assessment of lifelines, " presented at the 16th ASCE Eng
- D Reed
- N Nojima
- J Park
D. Reed, N. Nojima, and J. Park, " Performance assessment of lifelines, " presented at the 16th ASCE Eng. Mechanics Conf., Seattle, WA, 2003.
Negative bino-mial regression of electric power outages in hurricanes respectively. Currently, she is an Assistant Professor in the School of Civil and Environ-mental Engineering
- H Liu
- R Davidson
- D Rosowsky
- J Stedinger
H. Liu, R. Davidson, D. Rosowsky, and J. Stedinger, " Negative bino-mial regression of electric power outages in hurricanes, " J. Infrastruct. Syst., vol. 11, no. 4, pp. 258–267, Dec. 2005. Haibin Liu received the B.S. degree from Tsinghua University, Beijing, China, in 2000 and the M.S. and Ph.D. degrees in civil engineering from Cornell Uni-versity, Ithaca, NY, in 2003 and 2006, respectively. Rachel A. Davidson received the B.S.E. degree from Princeton University, Princeton, NJ, in 1993, and the M.S. and Ph.D. degrees in civil engineering from Stanford University, Palo Alto, CA, in 1994 and 1997 respectively. Currently, she is an Assistant Professor in the School of Civil and Environ-mental Engineering, Cornell University, Ithaca, NY. Tatiyana V. Apanasovich received the B.A. and M.S. degrees in economic cy-bernetics from Belarusian State University, Minsk, Belarus, in 1999 and the Ph.D. degree in statistics from Texas A&M University, College Station, in 2004. Currently, she is an Assistant Professor in the School of Operations Research and Industrial Engineering at Cornell University, Ithaca, NY.
Empirical estimation of lifeline outage time in seismic disaster
- N Nojima
- Y Ishikawa
- T Okumura
- M Sugito
N. Nojima, Y. Ishikawa, T. Okumura, and M. Sugito, "Empirical estimation of lifeline outage time in seismic disaster," in Proc. U.S.-Jpn.
Joint Workshop 3rd Grantee Mtg., U.S.-Jpn. Cooperative Res. Urban
Earthquake Dis. Mit., Seattle, WA, 2001, pp. 516-517.
Electric Utility Business Customer Satisfaction Study
- J D Power
J. D. Power, Electric Utility Business Customer Satisfaction Study,
2006, and Associates.
A computer program for the estimation of restoration times during storms
- L Cooper
- N Schulz
- T Nielsen
L. Cooper, N. Schulz, and T. Nielsen, "A computer program for the estimation of restoration times during storms," in Proc. 60th Amer. Power
Conf.: Reliability and Economy-Technology Focus for Competition
and Globalization, Chicago, IL, 1998, pp. 777-780, pt. 2, 2.
Optimization of post-earthquake restoration of city gas systems
- R Isoyama
- T Iwata
- T Watanabe
an empirical technique for diagnosing precipitation type from model output
- ramer
J. Ramer, "An empirical technique for diagnosing precipitation type
from model output," in Proc. 5th Int. Conf. Aviation Weather Systems,
Vienna, VA, 1993, pp. 227-230, Preprints.
lifeline interaction and post earthquake urban system reconstruction
- zhang
R. Zhang, "Lifeline interaction and post earthquake urban system reconstruction," in Proc. 10th World Conf. Earthquake Eng., Rotterdam,
Netherlands, 1992, pp. 5475-5480.
Using Weather Research and Forecasting Model Output to Estimate ice Accretion During Freezing Rain Events
- A Degaetano
- B Belcher
- P Spier
A. DeGaetano, B. Belcher, and P. Spier, Using Weather Research and
Forecasting Model Output to Estimate ice Accretion During Freezing
Rain Events, unpublished.
&Ccedil;a&gcircPost-earthquake restoration planning for Los Angeles electric power
- R Nan
- S Davidson
- Guikema
A model methodology for assessment of seismic vulnerability and impact distribution of water supply systems
A model methodology for assessment of seismic vulnerability and
impact distribution of water supply systems Appl. Technol. Council
(ATC), Redwood City, CA, Rep. ATC-25-1, 1992.
National institute of building sciences (NIBS)
"National institute of building sciences (NIBS)," HAZUS Tech.
Manual, 1997, Washington, D.C.
empirical estimation of lifeline outage time in seismic disaster
- nojima
optimization of post-earthquake restoration of city gas systems
- isoyama
a computer program for the estimation of restoration times during storms
- cooper