Quantifying system resilience using probabilistic risk assessment techniques

Resilience provides a framework to assess a system's likelihood to succeed in its mission even as disruptions perturb the operations. A system's resilience is therefore essentially a risk proposition of the mission succeeding and as such can be quantified using probabilistic risk assessment (PRA) techniques developed over the past three decades. Reliability engineering methods for evaluating hardware are insufficient by themselves, as they do not examine procedural mitigations, system margin, or human training applied to overcome anomalies. A resilient system needs to address failures, regardless of whether their cause is component malfunction, operator error, or external disruption, and continue to operate (perhaps at reduced functionality) within off-nominal operating environments. Resilience approaches look beyond hardware-only solutions to generate additional mitigation concepts to prevent, withstand, adapt to, and rapidly recover from failures or external disruptions. The methods and techniques used to produce PRAs encompass not only the hardware but also the operating procedures, the contingency plans, the software, and the physics of the eventual consequences. This article discusses how PRA is applied to quantifying system resilience and focuses on two aspects: scenario development and uncertainty quantification. © 2019 John Hopkins University. All rights reserved.