Uncertainty Quantification and Propagation in Linear and Nonlinear Dynamic Structural Systems.
Song, Mingming.
2019
-
This dissertation
is concerned with quantification and propagation of modeling uncertainties in the
application of finite element model updating to linear and nonlinear dynamic civil
structures. Civil structures are large-scale complex systems and behave nonlinearly in
nature. Effective characteristics of civil structures such as mass, stiffness and
damping can have significant variability ... read moredue to changing environmental, ambient and
operational conditions. A hierarchical Bayesian model updating approach has been applied
to quantify the variability of modeling parameters (mostly stiffness in considered
applications of this thesis) by estimating probability distribution (hyperparameters) of
the model parameters (e.g., mean and covariance of an underlying normal distribution).
It is shown that the estimated uncertainty converge to a constant level for the
hierarchical Bayesian method, compared to traditional Bayesian method in which parameter
uncertainty is reduced infinitely as more data is used. Dynamics structural response can
be affected by geometric and/or material nonlinearities. For structures with geometric
nonlinearity, a new model updating strategy is proposed using nonlinear normal modes
(NNMs). A Bayesian model updating approach using NNMs is proposed and applied to
calibrate a nonlinear cantilever beam with local nonlinearity. To perform model updating
of civil structures with material nonlinearity, an adaptive Kalman filter is proposed.
Kalman filters have received increased attention in civil engineering due to their
potential capability of online identification and structural health monitoring. Some
applications have shown their capability for parameter estimation, joint state-parameter
estimation or input estimation. Adaptive Kalman filter methods for nonlinear model
calibration are proposed and applied to numerical applications with more accurate and
robust parameter estimation observed.
Thesis (Ph.D.)--Tufts University, 2019.
Submitted to the Dept. of Civil Engineering.
Advisor: Babak Moaveni.
Committee: Gaetan Kerschen, Costas Papadimitriou, Masoud Sanayei, and Andreas Stavridis.
Keyword: Civil engineering.read less - ID:
- k643bd575
- To Cite:
- TARC Citation Guide EndNote
