Abstract
Dynamic deflection of a railroad sleeper works as an indicator of ballast stiffness, reflecting the health conditions of a ballast track. However, difficulty exists in measuring dynamic deflection of a railroad sleeper by conventional deflection transducers such as a linear variable differential transformer (LVDT) or a potentiometer. This is because a fixed reference point is unattainable due to ground vibrations during train passage. In this paper, a patented signal processing technique for evaluation of pseudo-deflection is presented to recover dynamic deflection of a railroad sleeper using a coupled measurement of acceleration and strain at the concrete sleeper. The presented technique combines high-frequency deflections calculated from double integration of acceleration and low-frequency deflections determined from strains. Validity of the combined deflections was shown by the deflections measured with a camera target on a concrete sleeper, captured by a high-resolution DSLR camera with superb video capturing features and processed by computer vision techniques, such as Canny edge detection and Blob analysis.
Highlights
Dynamic deflection of a railroad sleeper is a good indicator of the health condition of a ballast track
The sleeper deflections measured by linear variable differential transformer (LVDT)’s or potentiometers tend to be underestimated due to movement of a mounting rod induced by train loading
The individual image frames in the video were processed using two different computer vision (CV) techniques to determine the dynamic deflections of a sleeper
Summary
Dynamic deflection of a railroad sleeper is a good indicator of the health condition of a ballast track. Dynamic deflections of a railroad sleeper have been measured for more than several decades for various projects These include the development of an environmentally-friendly concrete sleepers [1], stiffness evaluation of a ballast track for a high-speed railway, and stability checks of an approach block between a bridge and a soil embankment. Other effective approaches of measuring sleeper deflections include piecewise integration of an acceleration history [9], and an LVDT mounted on the settlement pegs embedded into ballasts [10]. The individual image frames in the video were processed using two different computer vision (CV) techniques to determine the dynamic deflections of a sleeper. The CV techniques adopted for this validity test were Canny edge detection and Blob analysis
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call