STOCHASTIC SCOUR ASSESSMENT OF RAILWAY BRIDGE PIER UTILIZING MICROTREMORS

Abstract

 This study intends to propose stochastic methods for real-time scour assessment of a railway bridge pier during swollen river water periods using microtremors from accelerometers deployed on top of it. When the natural frequencies of the pier are identified using ambient microtremors, there is a large variation in the identification due to uncertainty. This study, thus, proposes three scour-warning indicators derived from cumulative density function, marginal probability distribution, and logistic function modeled by Levy Flight Distribution at a normal condition. They are applied to the observed data at a swollen river water period, and it is demonstrated that the probability of scour progress is low. Furthermore, the validity of indicators using observations that pseudo-replicated acceleration waveform during scour progress is figured out considering changes in indicators.