Survivability analysis of a sewage treatment facility using hybrid Petri nets

    Research output: Contribution to journalArticleAcademicpeer-review

    8 Citations (Scopus)
    2 Downloads (Pure)

    Abstract

    Waste water treatment facilities clean sewage water from households and industry in several cleaning steps. Such facilities are dimensioned to accommodate a maximum intake. However, in the case of very bad weather conditions or failures of system components, the system might not be able to accommodate all waste water. This paper models a real waste water treatment facility, situated in the city of Enschede, the Netherlands, with Hybrid Petri nets with general transitions, to analyse under which circumstances the existing infrastructure will overflow. Comparing to previous models an structural extension is proposed, and one limitation is tackled. First, we extended the hybrid Petri net formalism with guard arcs and dynamic continuous transitions, to be able to model dependencies on continuous places and the rates of continuous transitions. Secondly, we tackle the restriction of having only a single general transition, by proposing a new discretization method. We introduce to different discretization methods, and compare their efficiency in a complex case study. Using recently developed algorithms for model checking STL properties on hybrid Petri nets, the paper computes survivability measures that can be expressed using the path-based until operator. After computing measures for a wide range of parameters, we provide recommendations as to where the system can be improved to reduce the probability of overflow.
    Original languageEnglish
    Pages (from-to)36-56
    Number of pages21
    JournalPerformance evaluation
    Volume97
    DOIs
    Publication statusPublished - Mar 2016

    Keywords

    • n/a OA procedure

    Fingerprint

    Dive into the research topics of 'Survivability analysis of a sewage treatment facility using hybrid Petri nets'. Together they form a unique fingerprint.

    Cite this