Discovery of Hidden Vulnerabilities in Critical Infrastructures

Weather and climate are responsible for about two-thirds of all the drinking water system failures in coastal communities. In the case of complex systems, failures and mishaps are often due to multiple causes that are hard to foresee and prevent. Unfortunately, it is hard to figure out cascades of failures with traditional mathematical models, which are tailored to specific events and do not address the wide range of situations caused by hazards that could impact such systems. To this purpose, we analyze the feasibility of a software tool that can help communities to foresee the effects of severe weather and climate-induced stressors on water systems.
In this study, we describe a conceptual model that we used to design an intelligent knowledge based system. The intelligent system, formalized as an ontology and a set of rules, uncovers hidden vulnerabilities in critical infrastructures. This is done by using knowledge previously gathered from water systems experts, stakeholders, and users in other cases. It is worth noting that, in our proposal, knowledge is continuously enriched during operation through automatic collection of end user’s contributions after initialization.
An initial set of vulnerabilities was obtained from past risk assessments conducted on water systems in coastal communities and was used to derive the interactions occurring between hazards, water systems, and stakeholders. All the interactions were extracted from vulnerability assessments made during a NOAA funded experiment that used the same well-tested protocol during several different focus groups of experts. We used the initial set of known risks to model interactions between system components and failures. We then built a model to associate vulnerabilities to the components of the infrastructure, a step needed to enable estimates of the likelihood of reoccurrences of the scenarios. Next, a semantic based app uses the model and adapts the stored water system experiences to generate new failure scenarios that stakeholders can evaluate in the case of different systems.
It is anticipated this application can be valuable to assist experts in the analysis of risks that are hard to predict due to the large number of components that are normally part of a water system. Also, since the application can acquires new knowledge through on-line surveys, it can be used to facilitate decision makers and vulnerable communities to discover hidden sources of risks that need to be considered.