This paper mainly deals with the identification of structural modal parameters, from output accelerometers only, using a vector autoregressive moving average (VARMA(p,q)) model method. The problem of determining the order p of the AR part, or equivalently the number of modes in a frequency band, is also examined using a combination of the multivariate minimum description length (MDL) criterion and an overdetermined instrumental variable sequence. The AR coefficients are then obtained from this instrumental variable sequence. The companion matrix is formed and the modal parameters of the vibrating system deduced. Numerical and experimental results are treated to show the effectiveness of this new procedure. An experimental test of a car door is treated. A random and unmeasured excitation is applied on the car door and eight accelerometers are considered to estimate the modal parameters of this mechanical system. The sampling frequency is 400 Hz and 4096 data points are collected for each channel. The purpose is to determine the order of the multivariate AR part, or equivalently, the number of modes in the frequency band (0; 200 Hz), the eigenfrequencies and damping coefficients of this car door from output accelerometers only.


    Access

    Access via TIB

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    A new method for model order selection and modal parameter estimation in time domain


    Additional title:

    Ein neues Verfahren zur Auswahl der Modellordnung und Abschätzung der Modalparameter im Zeitbereich


    Contributors:
    Lardies, J. (author) / Larbi, N. (author)

    Published in:

    Publication date :

    2001


    Size :

    17 Seiten, 6 Bilder, 7 Tabellen, 13 Quellen




    Type of media :

    Article (Journal)


    Type of material :

    Print


    Language :

    English