System Identification in Structural Dynamics Using Blind Source Separation Techniques

Separation and identification of structural modes in largely underdetermined scenarios using frequency banding

“…The numerical examples have been introduced in Refs. [8] [29], while two sets of the experimental data come from a round-robin test organized on occasion of the IMAC XXVII 2011 conference [35].…”

“…This is an academic example already analyzed in Refs. [8,34]. The separation performance is first investigated with proportional damping, which has been theoretically shown to be favorable for BSS [12], and next with non-proportional damping, which is known to be more challenging.…”

“…It was closely followed by Refs. [8,9] where it was demonstrated that SOBI (second-order system identification), a second-order 2 BSS algorithm based on the crosscorrelation of the data, improved the performance of separation as compared to ICA. From there on, several research works have compared BSS algorithms and investigated their limitations in OMA applications [10,11].…”