Bayesian Updating and Model Class Selection for Hysteretic Structural Models Using Stochastic Simulation

Abstract

System identification of structures using their measured earthquake response can play a key role in structural health monitoring, structural control and improving performance-based design. Implementation using data from strong seismic shaking is complicated by the nonlinear hysteretic response of structures. Furthermore, this inverse problem is ill-conditioned1 for example, even if some components in the structure show substantial yielding, others will exhibit nearly elastic response, producing no information about their yielding behavior. Classical least-squares or maximum likelihood estimation will not work with a realistic class of hysteretic models because it will be unidentifiable based on the data. It is shown here that Bayesian updating and model class selection provide a powerful and rigorous approach to tackle this problem when implemented using a recently developed stochastic simulation algorithm called Transitional Markov Chain Monte Carlo. The updating and model class selection is performed on a previously-developed class of Masing hysteretic structural models that are relatively simple yet can give realistic responses to seismic loading. The theory for the Masing hysteretic models, and the theory used to perform the updating and model class selection, are presented and discussed. An illustrative example is given that uses simulated dynamic response data and shows the ability of the algorithm to identify hysteretic systems even when the class of models is unidentifiable based on the data.

Real-Time Estimation of Fault Rupture Extent Using Near-Source versus Far-Source Classification

“…However, the Bayesian model class selection automatically accounts for the trade-off between the complexity of the model (e.g., the number of parameters) and the fit of the data to find a well-balanced model (Beck and Yuen, 2004). A useful information-theoretic interpretation of this trade-off is given in Muto and Beck (2007).…”

Simulation-Based Probabilistic Tsunami Hazard Analysis: Empirical and Robust Hazard Predictions

“…p(T i ) is the prior distribution of the model T i and can be taken equal to 1/N. Finally, p(D|T i ) is called evidence for the model T i provided by the data D. According to Muto and Beck (2008), the log-evidence can be expressed as the difference of two terms:…”

On the Performance of Online Parameter Estimation Algorithms in Systems with Various Identifiability Properties

“…This issue of unidentifiability due to presence of noise and low excitations will be inherent in many real-life non-linear systems. In the study by Muto and Beck (2008), the example is given of non-linear hysteretic structures subjected to seismic event. If some parts of the structure exhibit only linear behavior during the event, no information on their yielding behavior (and associated parameters) will be available from the measurements.…”

“…In contrast, Bayesian techniques consider parameters as random variables (RVs) and make use of the measurements to update some prior knowledge (prior probability density function pdf, or moments) and thus yield the posterior pdf (or moments). A crucial advantage of the Bayesian framework lies in the fact that it is able to tackle ill-conditioned problems where some or all parameters cannot be uniquely identified based on available measurements (Muto and Beck, 2008), which is a major topic of this paper. For identification of parameters, Bayes' theorem can be written as…”