A prescriptive model to assess the socio-demographics impacts of resilience improvements on power networks
Introduction
The concept of resilience relates to the time evolution of a steady-state system after a disruption. In some domains (e.g., predator-prey populations), the system can transform to a new equilibrium state in response to a shock, and cannot necessarily return to its pre-shock state [[1], [2], [3], [4], [5], [6]]. However, for disaster resilience in urban settings, the emphasis is generally on the system's ability to return to its original operational state after a shock [5,7]. In order to identify whether a system returns to original or a new equilibrium, a quantitative system performance function formulation, also dubbed as figure of merit [3], is needed.
Depending on the infrastructure domain and the governing agency (e.g., department of transportation, power utility company), the system performance and functionality relate to various infrastructure parameters. The performance of transportation networks is related to network connectivity and travel times, whereas the performance of power networks is related to the number of customers with electricity. However, any urban settlement is a socio-technical environment where the infrastructure and society interact. Particularly for urban areas, the resilience refers to a bounce back to “normal” daily life, including all economic and leisure activities. Thus, the restoration of infrastructure does not tell the whole story of resilience. For instance, the status of schools is a critical resilience indicator [8,9] as for almost all parents, reopening of schools after a disaster means that daily life returning to “normal” so that they can get back to their daily routine [10]. Similarly, the functionality of public and private institutions (e.g., re-opening of public offices and businesses) also contributes to the return to “normalcy.”
Platt [6] also points out the ambiguity in “what is normal?” and argue that the recovery speed is not the only performance measure to consider unless “building back better” is achieved. In that respect, focusing purely on infrastructure recovery can overlook the societal impacts and conceal the true return to normality. In addition, the bounce-back to the normality is not necessarily experienced simultaneously by all segments of the population. The concept of “social vulnerability” addresses this concern by emphasizing different patterns of recovery within different population groups that live at the same hazard region [11]. Tapsell et al. [12] argue that the social impacts of hazard exposure often fall disproportionately on the most vulnerable populations such as the poor, minorities, children, the elderly and disabled. Vale [13] argues that resilience cannot remain useful as a concept unless it is associated with improvement of the disadvantaged groups’ life prospects. Vale [13] further discusses that this social dimension is often lost in conceptualizations of resilience in engineering and ecology domains.
The literature provides numerous conceptual frameworks which identify the factors to incorporate the technical and social considerations [[14], [15], [16], [17]] as well as economic factors [18] into disaster resilience and component importance assessment. Both for technical and social parameters, these studies commonly cite data availability to be an important challenge in order to implement and validate the proposed frameworks. The rare nature of the disasters exacerbates this problem since the disaster impact and response operations data are difficult to obtain and are mostly unique to each disaster.
Due to aforementioned data constraints, resilience improvements are performed within costly lessons-learned cycles after each disaster. At every cycle, the observed and some of the anticipated problems are fixed in an ad-hoc manner, just to wait for the next disaster to understand whether the solution or policy was effective, and/or was equitable across different population groups. In other words, the investments and policy decisions for resilience improvements are generally performed based on the after-the-fact descriptive analysis and do not necessarily include a prior assessment of the anticipated technical and societal impacts. On the other hand, rehearsing critical decisions before an event and applying science to inform decision making for building-back-better are emphasized as key issues for governments and decision makers [6]. In order to break the costly lessons-learned cycle and support improvement decisions, there is a need for prescriptive models that account for social vulnerability. Accordingly, the disaster preparedness plans can help prioritize resilience improvements based on the characteristics of the affected population simultaneously with infrastructure-related technical data.
This paper develops a prescriptive resilience model (to the extent possible with data availability) for one of the critical infrastructure networks, namely the power grid. Within this model, power infrastructure failure rate and recovery duration models are developed based on Hurricane Hermine power outage data obtained from the City of Tallahassee, FL. For the component failures, physical factors such as component type and age, and building age in the surrounding area are used. For the component restoration, the factors such as component age, critical facilities, and land use characteristics are considered. By using Monte Carlo simulation, multiple spatio-temporal infrastructure failure and recovery scenarios are created to represent different hurricane impact cases. The time evolution of recovery (i.e., percentage of population with power at any time) is broken into socio-economic categories such as income, age, and ethnicity.
The analyses provide the base case recovery findings with a focus on the affected socio-demographics for future hurricanes similar to Hermine. The infrastructure failure and recovery models help test the impacts of improvement alternatives. Two resilience policy/investment decisions are selected for further analysis: 1) investment to renew infrastructure components, and 2) increasing the response crew size for faster recovery. In addition, spatially specific improvement scenarios based on population income are also conducted. For each improvement scenario, the infrastructure failure and recovery rate models are used to simulate the improved recovery performance and recalculate the socio-demographic impacts. Overall, it is shown that the developed framework can calculate the varying impacts of improvement decisions on different demographics, and can help policy makers and emergency managers to prioritize resource investment plans based on affected demographics and maintain equity.
The paper is structured as follows: First, a literature review is provided and the lack of research on the impacts of infrastructure recovery actions on different populations segments is emphasized. Second, the Hurricane Hermine power outage data for the City of Tallahassee, FL are described, and the overall prescriptive modeling approach is explained. Third, the power network component failure and recovery models are described and estimated. These models help create power network resilience scenarios under hurricane threat, and help circumvent the problem of scarce data availability for assessment of resilience improvements. Then, the hypothetical scenarios are utilized to assess the socio-demographics impacts of alternative resilience improvement actions. Last, the conclusion and future research directions are summarized.
Section snippets
Literature review
With the ongoing climate change, extreme weather events such as hurricanes have become more prevalent with more devastating outcomes. Recently, three significant hurricanes, namely Hermine (2016), Irma (2017), and Michael (2018) hit the Florida Gulf Coast. The City of Tallahassee, capital of Florida, was affected severely by these hurricanes; particularly Hermine caused widespread power outages and roadway closures for a long duration. The impact of Hermine showed that the infrastructure of
Study area
In this study, Hurricane Hermine's impact on the power network of the City of Tallahassee, FL was analyzed. Tallahassee is the capital of Florida with a total population of 190,894 and is a home to two major universities, making it an important urban region in the Northwestern Florida. Hurricane Hermine hit Tallahassee on September 2nd, 2016. It was the first hurricane to make landfall in Florida since Hurricane Wilma in 2005, and was the first hurricane to directly hit Apalachee Bay since
Data
The power infrastructure data include municipality power lines (feeders) as well as components such as circuit breakers, reclosers, sectionalizers, switches, fuses, and transformers that serve 126,737 electricity customers in the city. These components are connected to each other within the power infrastructure, and all customers are connected to the power infrastructure to access power.
The power infrastructure is divided into feeder subnetworks and components connected to each feeder
The approach
The overall framework proposed in this study is presented in Fig. 3. The objective is to provide a modeling framework that can help decision makers assess resilience policy alternatives not only with respect to the technical aspects, but also with respect to the socio-demographic impacts. For this purpose, first, the available power network data is utilized to develop the component failure and power restoration models. By using these models, hypothetical hurricane scenarios are created through
Scenario generation for hurricane impacts on the power network
In order to generate scenarios for hurricane impacts on the power network, two models are estimated: 1) component failure and 2) component restoration. In essence, the failure model provides the spatial distribution of the power outages, and the restoration model provides the temporal recovery timelines that is necessary to calculate the resilience curves.
Assessment of alternative actions for resilience improvement
The resilience of urban systems is a multifaceted issue that is affected by external/exogenous factors such as the magnitude of an event as well as internal/endogenous features such as the condition of the infrastructure or preparedness of governments/communities. In general, the exogenous factors are outside the control of decision-makers and planner; however, the decision makers can act on altering the endogenous factors. Platt [6] studies earthquakes from multiple countries and the findings
Assessment of targeted policies for social equity
Overall, the developed framework provides a quantitative evaluation for strategies that are expected to yield improvements. For instance, the findings indicate that renewal of components more than 5-years old results in an average of 7 h earlier 90% recovery compared to reduction of recovery times by 10%. Meanwhile, the reduction of recovery times by 20% leads to 2 h earlier 90% recovery compared to the renewal of more than 5-years-old components. Such results establish the comparison of
Conclusion
This study develops a prescriptive modeling framework to assess the socio-demographics impacts of resilience improvement actions with a focus on power networks. The prescriptive nature of the approach help break the lessons-learned cycle in disaster management and help decision-makers have before-the-fact anticipation of the potential resilience improvement actions on socio-demographic groups. The framework utilizes the power outage and recovery data from the technical side, and demographics
Future work
It is mainly a matter of data availability to increase the accuracy of the model, to test more detailed scenarios, and to upgrade the framework itself towards a decision-making tool. First candidate for improvement is component failure model. By utilizing spatio-temporal wind speed and direction data (that is in enough detail to project on individual infrastructure components) can be used to formulate a component failure model that can be applied for hurricane scenario analysis. Nevertheless,
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgment
The authors would like to thank the City of Tallahassee, especially Mr. Michael Ohlsen, for providing data and valuable insight. The contents of this paper and discussion represent the authors' opinion and do not reflect the official view of the City of Tallahassee. This material is based upon work supported by the National Science Foundation under Grant No. 1737483 and Grant No. 1640587.
References (74)
Organizational learning for resilient cities, through realizing eco-cultural innovations
J. Clean. Prod.
(2013)- et al.
Generic metrics and quantitative approaches for system resilience as a function of time
Reliab. Eng. Syst. Saf.
(2012) - et al.
An economic framework for the development of a resilience index for business recovery
Int. J. Disaster Risk Reduct.
(2013) - et al.
An inclusive and adaptive framework for measuring social resilience to disasters
Int. J. Disaster Risk Reduct.
(2018) - et al.(2016)
Network structure and travel patterns: explaining the geographical disparities of road network vulnerability
J. Transport Geogr.
(2009)- et al.
Road network vulnerability analysis of area-covering disruptions: a grid-based approach with case study
Transport. Res. Pol. Pract.
(2012) - et al.
Efficiency of scale-free networks: error and attack tolerance
- et al.
Optimizing power system investments and resilience against attacks
Reliab. Eng. Syst. Saf.
(2017) - et al.(2014)
Logistic regression model for predicting the failure probability of a landslide dam
Eng. Geol.
International Journal of Disaster Risk Reduction Emergency evacuation planning policy for carless and vulnerable populations in the United States and United Kingdom
Resilience: a bridging concept or a dead end?
Plann. Theor. Pract.
Resilience and stability of ecological systems
Annu. Rev. Ecol. Systemat.
Factors affecting the speed and quality of post-disaster recovery and resilience
Resilience in the built environment
Build. Res. Inf.
Concepts and approaches of school centered disaster resilient communities
Promoting Community Resilience in Disasters
Lessons from crisis recovery in schools: how hurricanes impacted schools, families and the community
Sch. Leader. Manag.
Social vulnerability to environmental hazards
Soc. Sci. Q.
Social Vulnerability to Natural Hazards. CapHaz-Net WP4. Flood Hazard Research Centre – FHRC
The Politics of Resilient Cities: whose resilience and whose city?
Build. Res. Inf.
From fatalism to resilience: reducing disaster impacts through systematic investments
Disasters
Community disaster resilience: a systematic review on assessment models and tools
PLOS Curr. Disaster
Evaluating and visualizing the economic impact of commercial districts due to an electric power network disruption
Risk Anal.
National Hurricane Center Tropical Cyclone Report - Hurricane Hermine
Resilience thinking: integrating resilience, adaptability and transformability
Ecol. Soc.
Resilience benchmarking & monitoring review
A framework to quantitatively assess and enhance the seismic resilience of communities
Earthq. Spectra
Emergence of Scaling in Random Networks. Science (New York, N.Y.) 286(5439)
Catastrophic cascade of failures in interdependent networks
Nature
Network capacity reliability analysis considering traffic regulation after a major disaster
Cyber–physical system security for the electric power grid
Proc. IEEE
Statistical mechanics of complex networks
Rev. Mod. Phys.
Mean-field theory for scale-free random networks
Cited by (5)
The human factor in the disaster resilience modeling of critical infrastructure systems
2023, Reliability Engineering and System SafetyPost-hurricane vegetative debris assessment using spectral indices derived from satellite imagery
2021, Transportation Research RecordCity transportation network vulnerability to disasters: the case of Hurricane Hermine in Florida
2021, Environmental Hazards
- 1
Current Affiliation: Department of Civil Engineering, University of Twente, Enschede, The Netherlands