A New Model Updating Procedure for Reliability-Based Damage and Load Identification of Railway Bridges

Abstract

In the current work, a new reliability-based method is presented for damage and load identification of railway bridge structures using finite element model updating in the presence of uncertainty. The bridge structure is modelled as an Euler-Bernoulli beam and the train is modelled as a series of axles in the form of sprung-mass with an unknown weight. Since the bridge-vehicle system is time-variant, the finite element model updating procedure is used as a response-based method. The effect of model and measurement uncertainties on identificationresults is investigated. The efficiency of the reliability-based method is compared with the deterministic and traditional probabilistic methods. In addition, the effect of critical parameters such as damping, speed and mass ratio on the accuracy of the proposed method is also discussed. It is observed in numerical models that the reliability-based method is the most accurate method for simultaneous identification of railway bridges in the presence of uncertainty. In addition to the damage index (DI), which indicates the location and the extent of damaged elements, the proposed method provides the probability and reliability of identified damages (RI) as well as the extent of unknown moving loads accurately.