On Evaluating System Resilience by the Degree of Order Disruption

Abstract

Resilience curves describe the diminished performance of a system over time following a shock, with the area integral above the curve representing the persistence of the lost performance. For example, in a power system, the lost power generation is measured in megawatts, and the persistence of the disruption is in megawatt hours; in a communications system, the performance in gigabits per second can be integrated over time as gigabits. The persistence metrics from the examples are not comparable. Ongoing engineering systems work is focused on disrupted order rather than disrupted performance. The Kendall's tau statistic is used to compare an initial, pre‐shock set of priorities with the priorities after the shock. This paper will describe complementing system performance with system order in the estimation and integration of resilience curves, thus enabling a new quantification of resilience as a disruption of system order can be used across application domains of risk analysis.