FRF-based probabilistic modal parameter identification of structures with known seismic input
Küçük Resim Yok
Tarih
2023
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Academic Press Ltd- Elsevier Science Ltd
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
In this study, an FRF based Bayesian approach is presented for modal identification of civil structures with known seismic input. The proposed method utilizes the statistical properties of frequency response function (FRF) of the observed structure which is obtained from the accel-eration response (output) and the seismic input recorded at the base level. In this context, first, the joint pdf of prediction error term and seismic input is derived. Then, applying a linear vector transformation, the output term in the derived joint pdf is defined as depending on the FRF and the seismic input. Finally, the marginal probability of FRF is obtained as a complex ratio distri-bution from the transformed joint pdf. The posterior probability function for modal parameters is obtained by using Bayes' theorem. The modal parameters are obtained by maximum likelihood estimation of the constructed probability density function. Finally, the posterior uncertainties for modal parameters are calculated. Two numerical and two real data examples are presented to illustrate the computational efficiency of the proposed methodology against input-output and output only techniques.
Açıklama
Anahtar Kelimeler
Seismic SHM, Bayes? theorem, Uncertainty quantification, FRF based modal identification, Maximum-Likelihood Identification, Frequency-Domain, Uncertainty Intervals, Algorithm, Posterior
Kaynak
Mechanical Systems and Signal Processing
WoS Q Değeri
Q1
Scopus Q Değeri
Q1
Cilt
189
