Bridge Resistance Updating Based on the General Particle Simulation Algorithms of Complex Bayesian Formulas

Abstract

The existing bridges are subjected to time-variant loading and resistance degradation processes. How to update resistance probability distribution functions with resistance degradation model and proof load effects has become one of the research hotspots in bridge engineering field. To solve with the above issue, this paper proposed the general particle simulation algorithms of complex Bayesian formulas for bridge resistance updating. Firstly, the complex Bayesian formulas for updating resistance probability model are built. For overcoming the difficulty for the analytic calculation of complex Bayesian formulas, the general particle simulation methods are provided to obtain the particles of complex Bayesian formulas; then, with the improved expectation maximization optimization algorithm obtained with the combination of K-MEANS algorithm and Expectation Maximization (EM) algorithm, the above simulated particles can be used to estimate the posteriori probability density functions of resistance probability model; finally, a numerical example is provided to illustrate the feasibility and application of the proposed algorithms.