The basic services that make modern life possible are often taken for granted. But when the lights go out, the tap runs dry and the internet quits working, the vulnerabilities in infrastructure are quickly laid bare. Recovery can take weeks, and the poor are disproportionately affected.
Few realize how interconnected infrastructure systems are until they fail. A breakdown in one system can quickly ripple to others.
To keep services running in the wake of disaster, we need a new breed of professional who can work across disciplines.
Against this backdrop, Clemson University is offering a graduate research program in Resilient Infrastructure Systems to respond to industry’s urgent need for professionals who can:
- assess technological and societal risks,
- communicate those risks to decision makers, and
- devise strategies that improve community resilience.
Our program is backed by a $3-million grant from the National Science Foundation’s Research Traineeship Program, the first of its kind in South Carolina. Our focus is identifying and mitigating the vulnerabilities of complex, critical, and interdependent infrastructure systems. The need has grown measurably more urgent in the last 50 years and is only expected to heighten.
We have seen a ten-fold increase in the frequency and impact of weather events, and a significantly higher risk of man-made disasters since 9/11. Hurricane Katrina’s devastation to the Gulf Coast underscores the far-reaching effects of disaster. Six months after the storm passed, 46 percent of the affected oil facilities were still shut down, resulting in an estimated $15 billion in losses to the energy industry in 2005, not including the costs for restoration and recovery.
The extensive, profound changes caused by urbanization and globalization over the last few decades further complicate the problem. An infrastructure disruption in one area no longer affects only the local populace. It can have far-reaching consequences, thanks to the technology and supply chains that connect the globe. Cities are expected to become increasingly interconnected and larger. By 2050, 66 percent of the world’s population will live in urban areas, an increase of 12 percent, according to a 2014 estimate from the United Nations.
This is a new type of threat for the 21st century.
What is worse, infrastructure vulnerabilities tend to have more significant impacts in underserved regions. During Hurricane Katrina, many citizens from marginalized neighborhoods were trapped in flooded areas because they lacked means to evacuate. While the industrial impacts of Katrina are fully remedied, the societal impacts were still reverberating 11 years later, particularly in economically disadvantaged areas, such as the 9th Ward of New Orleans. The number of homeless increased by nearly a factor of four following Katrina.
For our society to prosper, we must train professionals who can conceptualize complex systems in which physical, cyber, and human infrastructure elements converge and transform this conceptual understanding into reliable computational models that are validated by data.
Just as importantly, these graduates must be equipped with skills to effectively communicate with their peers in other disciplines and with decision-makers to ensure cohesion between science and policy. Such skills are valuable, rare, and timely. The graduate research program in Resilient Infrastructure Systems is inspired by a vision to address these needs through:
- Curriculum development. Our program is a responsive, replicable, modular training program that synthesizes resources from across the campus to address the growing need for transdisciplinary system thinkers, model engineers, and data scientists;
- Transformation in graduate education. We embrace an educational research philosophy that targets both students and faculty to promote collaborative research communities with strong peer-learning aspects, addressing critical challenges in graduate education as identified by the U.S. government;
- Research with societal impact. Our program is a uniquely integrated research, training, and outreach program that studies infrastructure vulnerabilities that induce low-income regions to be disproportionately impacted by disasters, such as the White House’s Promise Zone along Interstate 95 corridor in South Carolina.
The core research goal of this project is to identify and mitigate infrastructure vulnerabilities through the use of advanced modeling and simulation. The problem is complex and requires us to account for interdependencies between each infrastructure component with uncertain feedback mechanisms.
For this, integrating mechanistic models with data-driven models of various fidelity becomes necessary to capture the complexity of physical, natural, cyber, and socio-economical systems across temporal and spatial scales. During model integration, modeling and simulation challenges arise from the need to:
- Identify key processes that combine to produce feedbacks and emergent system behaviors;
- Integrate models with vastly different ontologies, operating at different spatial and temporal scales; and
- Capture the full degree of detail and complexity required for representing these systems given limits on computational and other resources.
In addition, using model predictions in support of decision-making requires undertaking the challenging task of validating the integrated models against observable data and quantifying the uncertainties in the integrated model predictions.
The importance of these activities, widely known as model validation and uncertainty quantification, has been acknowledged by many government agencies including NASA, the Department of Energy, the Department of Defense, the Environmental Protection Agency and the National Science Foundation.
Beyond these technical challenges, we also recognize a need to incorporate decision-makers within the modeling process. Better models should mean better decisions, but this is not always the case. We must therefore account for the differences in human decision-making as an integral part of our modeling and simulation capabilities. To respond to these needs, we will pursue three research thrusts. We will:
- Develop integrated computational representations of interdependent infrastructure systems (which have traditionally been analyzed in silos) considering complex, poorly-understood feedback mechanisms and processes;
- Produce computer models informed by heterogeneous and high-volume data from multiple physical, ecological, engineered, and social sources (ideally in real-time) to mitigate errors and uncertainties in the models and achieve situation-relevant, decision-quality (real-time) predictions; and
- Develop frameworks and strategies to enhance communication between model developer and decision-maker, thus enabling stakeholders to use model information in the most effective manner.