IL CORSO SI TERRA' NEL PERIODO: NOVEMBRE-DICEMBRE


The scope is to give a wide overview of the huge number of techniques which have been developed since the explosion
of modal analysis approach, refining and fitting the identification procedures to be adapt to ouput-only, non-linear and time variant
behaviours.

Tecniche di identificazione dei sistemi di diagnostica
Il corso ha lo scopo di illustrare le più recenti tecniche di identificazione di sistemi dinamici a fini diagnostici e di monitoraggio.
Sono affrontate velocemente le tecniche lineari, ben conosciute, mentre è dato più rilievo alle tecniche per i sistemi non-lineari e quelli tempo-varianti, recentemente diventati di grande interesse per la comunità scientifica.
Sono in particolare evidenziate le tecniche dette output-only, senza misura dell'ingresso.
Metodi visti sono, in ambito lineare, ARMA-ARMAV, SSI, ITD-RDT e altro.
In ambito non lineare e tempo variante, sono considerate le tecniche semplici, come coherence function, superposition method, restoring force mapping, Hilbert Transform e più recenti metodi NL-SS e TV-SSI, oltre ai metodi TARMAV e le trasformate Wavalet e tempo-frequenza.

Systems Identification techniques for diagnostics”

The course is aimed to give a general overview on the dynamic system identification techniques, for linear, non-linear and time-varying systems. Participants should have a sound basis on signal processing and modal analysis techniques.
Linear systems are treated to resume and highlight the drawback and advantages, whilst the focus is more on non linear and,
in particular, time variant systems, recently widely adopted for modelling true engineering problems. An overview is given
on ARMA-ARMAV, SSI, ITD-RDT, and so on.
Wavelet technique is also treated, as other time-frequency methods; mainly, output-only methods are mainly considered.
Non-linear systems are treated by an historical point of view (such as coherence function, superposition method, restoring
force mapping, Hilbert Transform) and by treating the most recent techniques.
In order to qualifying and quantifying the non-linearities, the methods of the Restoring Force, the Wavelet Transform, the
Wiener and Volterra series, the NARMAX models, the Neural Nets and the NL-SSI are shown.
Time variant system are treated on the basis of the most recent development, with TARMA approach, TV-SSI, wavelet, and other
methods.