A feasibility study on void detection of cement-emulsified asphalt mortar for slab track system utilizing measured vibration data. (15th October 2021)

• Record Type: Journal Article
• Authors: Hu, Q.; Shen, Y.J.; Zhu, H.P.; Lam, H.F.; Adeagbo, M.O.
• Is Part Of: Engineering structures. Volume 245(2021)
• Publication Date: 2021-10-15
• Subjects: Bayesian model updating -- Bayesian model class selection -- Void detection -- CA mortar -- Ballastless slab track; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2021.112349
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: The newly developed two-phase model class selection algorithm for void detection is firstly proposed. The time-domain Bayesian model updating is successfully conducted for the CA void detection of the ballastless slab track by determining the most probable values (MPVs) of mortar stiffness scaling factors. Not only the void location but also the severity of CA mortar are detected for ballastless slab track. Abstract: This paper reports the feasibility study on the use of measured vibration of a railway slab track system in detecting void on the cement-emulsified asphalt (CA) mortar layer utilizing the Bayesian approach. By following the specification of the China Railway Track System (CRTS)-I ballastless slab track structure, two scaled models (with and without void in the CA mortar layer) were built and tested in the laboratory to demonstrate and verify the proposed CA void detection methodology. A three-dimensional finite element model was built using ABAQUS to calculate the time-domain data of the ballastless track system for Bayesian model class selection and model updating. A new two-phase model class selection algorithm was developed for identifying the CA void region. The proposed methodology identified the distribution of CA mortar stiffness using impact hammer test data from the scaled CRTS-I ballastless slab panel under laboratory conditions. The model updating results are consistent with the simulated CA mortar stiffness distribution. In addition, the posterior uncertainties of the identified CA mortar stiffness under different sensor configurations were quantitatively investigated. The results from the experimental case studies show that the proposed Bayesian methodology is feasible to detect CA void even with only a single accelerometer with the associated posterior uncertainties being kept at an acceptable level.