Resilience Quantification of Multi-Component Systems under Interdependent Multi-Hazard

Abstract

Engineered systems and critical infrastructures (CIs) have been rapidly developed worldwide throughout the last several decades. These systems usually consist of components with heterogeneous configuration, functionality and reliability metrics. Throughout an engineered system’s desired life, its components are subjected to a wide range of unexpected hazards; these hazards occur randomly (in terms of frequency and severity) and meanwhile, dependently. Though recovery measures commence upon these components’ failure, hazards may recur or exhibit a dynamic hazard rate, resulting in a non-monotonic recovery process. Worse, components exhibit stochastic/economic/structural dependency. The resilience of such systems needs to be carefully checked to assess system’s capacity to maintain a high performance and recover rapidly.

We herein quantitatively assess the resilience of engineering systems whose components are subjected to heterogeneous hazards and function dependently. First, for an arbitrarily independent component, we investigate the whole spectrum of the hazards that pose threats to its functionality. Specifically, the interdependency between hazards is categorized into two sorts: frequency acceleration or severity aggravation. We then derive the analytical expressions of the component’s recovery process, along with its availability over time. Then, by integrating all components, system’s reliability (i.e., performance) can be obtained as a multimodal function. A stage-dependent resilience quantification is then developed to incorporate value and changing ratio of system’s performance.