Abstract:The estimation of the <a class="topic-link" href="/topics/engineering/posterior-probability">posterior probability</a> distribution of unknown parameters remains a challenging issue for model updating with uncertainties. Most current studies are based on <a class="topic-link" href="/topics/computer-science/stochastic-simulation">stochastic simulation</a> techniques. This paper proposes a novel variational <a class="topic-link" href="/topics/mathematics/bayesian-inference">Bayesian inference</a> approach to estimate posterior probability distributions by using the vibration responses of <a class="topic-link" href="/topics/engineering/civil-engineering-structure">civil engineering structures</a>. An adaptive Gaussian process modeling technique is used to represent the "expensive-to-evaluate" likelihood function, and the unknown posterior probability distribution is represented using a <a class="topic-link" href="/topics/engineering/gaussian-mixture-model">Gaussian mixture model</a>. The evidence lower bound (ELBO) and its gradients can be computed analytically using the built Gaussian process and mixture models. The unknown parameters in the Gaussian mixture model can be identified by maximizing the value of ELBO. The stochastic gradient descent method is applied to perform the optimization. Numerical studies on an eight-story shear-type building and a <a class="topic-link" href="/topics/engineering/simply-supported-beam">simply supported beam</a> are conducted to verify the accuracy and efficiency of using the proposed approach for probabilistic model updating and damage identification. Experimental studies on a laboratory steel frame structure are also conducted to validate the proposed approach. Results demonstrate that the posterior probability distributions of the unknown structural parameters can be successfully identified, and reliable probabilistic model updating and damage identification can be achieved.

Multiresolution Bayesian nonparametric general regression for structural model updating

Bayesian model updating of a 20-story office building utilizing operational modal analysis results

Structural modal identification and MCMC-based model updating by a Bayesian approach

A Bayesian neural network approach for probabilistic model updating using incomplete modal data

Bayesian Structural Model Updating with Multimodal Variational Autoencoder

Markov Chain Monte Carlo-based Bayesian method for nonlinear stochastic model updating

Bayesian Forecasting Approach for Structure Response Prediction and Load Effect Separation of a Revolving Auditorium

Bayesian Structural Model Updating Using Ambient Vibration Data Collected by Multiple Setups

Efficient model updating and health monitoring methodology using incomplete modal data without mode matching

Vibration-based Bayesian model updating of civil engineering structures applying Gaussian process metamodel

Active learning structural model updating of a multisensory system based on Kriging method and Bayesian inference

An ensemble learning based Bayesian model updating approach for structural damage identification

Bayesian Model Updating Using Hybrid Monte Carlo Simulation with Application to Structural Dynamic Models with Many Uncertain Parameters

Design of multi-hidden-layer Bayesian neural networks for model updating

Bayesian Model Updating of a 250 M Super-Tall Building Utilizing an Enhanced Markov Chain Monte Carlo Simulation Algorithm

Bayesian Model Updating of a Simply-Supported Truss Bridge Based on Dynamic Responses

Developing efficient model updating approaches for different structural complexity- an ensemble learning and uncertainty quantifications

Probabilistic model updating via variational Bayesian inference and adaptive Gaussian process modeling

An efficient algorithm for architecture design of Bayesian neural network in structural model updating

A variational Bayesian inference technique for model updating of structural systems with unknown noise statistics

New Model Variable Index for Updating a High-Rise Building Structural Model Based on an Artificial Neural Network