This Critical Resilient Interdependent Infrastructure Systems and Processes (CRISP) project will identify new strategies to increase resilience in interdependent electric power, communication and natural gas networks. These three critical systems increasingly depend on one another to keep our energy and communication systems running. In some ways connections between these systems can make them work better, but in other ways connections can increase the chance of disastrous failures that could leave millions of people without heat, electricity or the ability to communicate. For example, a severe winter storm in the Northeastern United States could lead to both power grid failures and natural gas failures, leading to failures in telephone and Internet services, making it even more difficult to restore these critical services. Such "cascading failures" make it even harder for these systems to recover from natural disasters and intentional attacks. This project will identify strategies to make interdependent infrastructure systems more resilient to these cascading failures. Four Research Directions will combine to address this problem. Research Direction 1 will adapt new computational algorithms, such as Influence Graphs that can identify non-obvious critical connections and the Random Chemistry algorithm that can rapidly find critical triggering events, to the particular problems of cascading failures in interdependent infrastructure systems. Research Direction 2 will create new models of interdependence among natural gas, electric power and communication networks, which will form a testbed for computational algorithms. The resulting models will balance computational complexity and engineering detail by using detailed dynamical models of each system when necessary and simplified mathematical models when abstractions can be validated from real data. Research Direction 3 will develop and evaluate engineering solutions and coordination strategies that can mitigate harmful interdependencies and leverage beneficial interconnections. These will leverage insights from the application of new computational algorithms to the interdependence testbed, such as the identification of critical failure paths, to develop both real-time dynamic rescheduling algorithms and cost-effective long-term planning strategies. Research Direction 4 will use stakeholder interviews to evaluate the diverse ways that the electricity, natural gas, and communications industries understand risk, and facilitate discussion among key industry participants regarding interdependencies among these systems. The results will reveal the most effective paths to integrating new control and planning strategies to increase resilience in these diverse systems.

This project will create significant societal benefits by uncovering new ways to reduce the risk of catastrophic failures among critical infrastructure systems. Because of interdependence among infrastructures, low probability, high cost cascading failures, which can have billions of dollars of economic and societal impacts, can contribute more to overall risk, relative to more frequent, small events. Reducing this risk can have enormous benefits to society. To ensure that results from this project have practical impacts the team will be guided by a Research Advisory Board that includes a large power grid operator (ISO New England), a software vendor for the electricity industry (GE/Alstom), a natural gas company (Vermont Gas), and the MITRE corporation. Furthermore, the project will integrate education and research through new curriculum and outreach to high school students. Public data that result from this project will be released through the github repository at: https://github.com/phines/infrastructure-risk, as well as through the project web site at www.uvm.edu/~tesla/project/nsf-crisp/. All research data associated with this project, including public and non-public data, will be preserved for at least 5 years after the end of the project.

Agency
National Science Foundation (NSF)
Institute
Division of Computer and Network Systems (CNS)
Type
Standard Grant (Standard)
Application #
1735463
Program Officer
Samee Khan
Project Start
Project End
Budget Start
2018-01-01
Budget End
2021-12-31
Support Year
Fiscal Year
2017
Total Cost
$1,199,898
Indirect Cost
Name
Massachusetts Institute of Technology
Department
Type
DUNS #
City
Cambridge
State
MA
Country
United States
Zip Code
02139