Identification of Time Varying Civil Engineering Structures using Multivariate Recursive Time Domain Models

P. Andersen, P. S. Skjærbæk, Poul Henning Kirkegaard

    Publikation: Bog/antologi/afhandling/rapportRapportForskningpeer review

    481 Downloads (Pure)

    Abstract

    This paper presents system identification results of an RC-structure using multivariable autoregressive moving-average vector models (ARMAV) on a running window, reursive ARMAW models (RARMAV) and Kalman filtering of a simple spring-mass-dashpot system. The identification techniques are used on simulated data generated by the non-linear finite element program SARCOF modelling a 6-storey 2-bay concete structure subjected to amplitude modulated Gaussian white noise filtered through a Kanai-Tajimi filter. The estiated natural frequencies and mode shapes of the two first modes ae compared with the smoothed quanties which have been obtained from SARCOF. The results show the usefulness of the technique for identification of a time varying civil engineering structure. It is found that all the techniques give reliable estiates of the frequencies of the two lowest modes and the first mode shape. Only the Kalman filtering gives reliable estimates of the second mode shape.
    OriginalsprogEngelsk
    UdgivelsesstedAalborg
    ForlagDept. of Building Technology and Structural Engineering, Aalborg University
    Antal sider12
    StatusUdgivet - 1996
    NavnFracture and Dynamics
    Nummer79
    Vol/bindR9619
    ISSN1395-7953

    Bibliografisk note

    Presented at ISMA21, Leuven, September 1996

    PDF for print: 12 pp

    Emneord

    • Spring-mass-dashpot system
    • RC-Structure
    • ARMAV

    Citationsformater