Infrastructure resilience considers how systems prepare for, respond to, and recover from disruptions. Within this broad topic, analytical interests and techniques vary; scenario-based analyses are those that specifically explore the progression of system states before, during, and after a disruptive event. This progression is commonly illustrated as a resilience curve: states are mapped to a performance measure and a summary metric quantifies overall system behavior. The resilience curve paradigm has been adopted across infrastructure domains (e.g., electrical, transportation, water) and analytical goals (e.g., evaluating contingency plans, optimizing recovery, comparing configurations). Motivated by extreme and ahistorical events, such analyses rely on modeling and simulation. Modelers are thus faced with the challenge of representing complex interdependencies, atypical situations, and stakeholder values within a finite system boundary.
This work contributes to the theory and practice of scenario-based resilience analyses across three chapters. The first chapter examines performance measures and summary metrics in a review of 273 publications. The review establishes taxonomies and synthesizes recommendations for their applicability and implementation. Additionally, this chapter highlights a common flaw in summarizing ensembles of curves and recommends further consideration of endogenous performance targets, thresholds, and weighting schemes. The second chapter explores the impact of time-varying context on quantitative assessments. A value weighting function is introduced to represent stakeholders’ perceived value of performance; correlation is described through a stochastic time offset. This chapter illustrates how common, performance-based metrics may be misleading. Results establish hazard categories for value-weighted analyses and suggest methods to enhance resilience beyond improving infrastructure performance. The third chapter proposes a novel modeling framework: Multiverse Simulation. In contrast to probabilistic techniques, this approach proactively explores possibilities for stakeholder insights and iterative model development. Through modified discrete event simulation, event counterfactuals establish diverging timelines; stakeholders are engaged to contemplate qualitatively-distinct system behaviors. Multiverse Simulation is shown to address six specified challenges such that it complements—but does not replace—existing modeling techniques.
About Craig Poulin
Craig Poulin is a PhD candidate in the Interdisciplinary Engineering program at Northeastern University under the advisement of Prof. Michel Kane. His research focuses on techniques to simulate infrastructure disruptions such that results provide practical resilience recommendations. Craig is an Air Force civil engineer officer with thirteen years of experience, and he currently serves as a Military Fellow with the Energy Systems Group at MIT Lincoln Laboratory. He received his B.S. degree in Electrical Engineering from Rensselaer Polytechnic Institute and his M.S. degree in Electrical Engineering from the University of Wisconsin- Madison. He is a registered Professional Engineer in the State of Ohio.
The conference will take place on Friday, June 11, 2021, from 2:00 pm to 3:00 pm. Join here.