Title: Cascade effects in critical infrastructure : predicting failure from flood events in interdependent infrastructure networks
Abstract: Interdependency within and between critical infrastructure networks increases their vulnerability
to failure after a natural hazard such as a
ood event. When operation of infrastructure assets gets
disrupted this can trigger failure in other infrastructure assets. This process is called a cascade
e�ect and can happen recursively which can cause initially small infrastructure disruptions to have
widespread consequences.
This study aims to predict cascade failure occurrence due to
oods in a selected set of infrastructure
networks at a detailed spatial scale. Using a given inundation map to assess direct failure of
assets, interdependencies between them are used to simulate indirect failure, i.e. assets failing due
to a cascade e�ect. Failure is described using a topology-based simulation model with aboveground
infrastructure assets represented as nodes and interdependencies between them as edges.
The modeling methodology is applied for the electrical, telecoms, gas and transportation networks
in the Dutch region Zeeuws-Vlaanderen. However, the aim is to devise a method which is
generically applicable both in other locations and other types of networks. In order to assess model
validity and determine potential areas of improvement, model results and premises are discussed in
an expert elicitation process. Operators of selected infrastructure networks are asked to comment
on di�erences between simulated and realistic failure behavior that ensue from modeling choices.
In the case study, failure occurs mostly around inundated areas, with direct failure generally
accounting for the larger share. This is attributed to key assets in selected networks not being
vulnerable to
ooding due to their geographical location, but also to the absence of higher order
networks in the case study. Indirect failure mostly occurs from intra-sectoral cascade e�ects, so
interdependence between di�erent infrastructure networks is not a driving force behind widespread
failure. Vulnerability to cascade e�ects can be reduced by introducing more network redundancy.
While this modeling methodology attempts to be generically applicable, di�erences between
infrastructure networks are encountered that require custom-�t modeling approaches. More information
speci�c to locations and networks can be introduced, but this does institute a need
for additional assumptions and data which is often unavailable. The currently applied modeling
methodology generally performs well in determining asset functionality during
ood events, especially
in networks with a clearly de�ned function and network commodity. However, it falls short
for more complex analyses as these require more network- and location-speci�c modeling. The
largest sources of inaccuracy are the premise that no network
ow is modeled and the connection
with infrastructure networks of higher order, such as the high voltage and national gas networks.
Publication Year: 2018
Publication Date: 2018-01-01
Language: en
Type: dissertation
Access and Citation
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot