A robust procedure for damage detection from strain measurements based on principal component analysis

A. Guemes, J. Sierra-Pérez, J. Rodellar, L. Mujica

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    2 Scopus citations

    Abstract

    FBGs are excellent strain sensors, because of its low size and multiplexing capability. Tens to hundred of sensors may be embedded into a structure, as it has already been demonstrated. Nevertheless, they only afford strain measurements at local points, so unless the damage affects the strain readings in a distinguishable manner, damage will go undetected. This paper show the experimental results obtained on the wing of a UAV, instrumented with 32 FBGs, before and after small damages were introduced. The PCA algorithm was able to distinguish the damage cases, even for small cracks. Principal Component Analysis (PCA) is a technique of multivariable analysis to reduce a complex data set to a lower dimension and reveal some hidden patterns that underlie.

    Original languageEnglish
    Title of host publicationStructural Health Monitoring
    Subtitle of host publicationResearch and Applications
    PublisherTrans Tech Publications Ltd
    Pages128-138
    Number of pages11
    ISBN (Print)9783037857151
    DOIs
    StatePublished - 2013
    Event4th Asia-Pacific Workshop on Structural Health Monitoring - Melbourne, VIC, Australia
    Duration: 5 Dec 20127 Dec 2012

    Publication series

    NameKey Engineering Materials
    Volume558
    ISSN (Print)1013-9826
    ISSN (Electronic)1662-9795

    Conference

    Conference4th Asia-Pacific Workshop on Structural Health Monitoring
    Country/TerritoryAustralia
    CityMelbourne, VIC
    Period5/12/127/12/12

    Keywords

    • Fiber optic sensors
    • Principal Component Analysis

    Fingerprint

    Dive into the research topics of 'A robust procedure for damage detection from strain measurements based on principal component analysis'. Together they form a unique fingerprint.

    Cite this