Train wheel damage detection based on deep learning

Conference Paper (2020)
Author(s)

Wen-Jun Cao (TU Delft - Resources & Recycling)

Research Group
Resources & Recycling
Copyright
© 2020 Wen-Jun Cao
More Info
expand_more
Publication Year
2020
Language
English
Copyright
© 2020 Wen-Jun Cao
Research Group
Resources & Recycling
Pages (from-to)
60-62
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

The train wheel flat is one of the most common damages in the railway system. It occurs when a wheel locks up while the train is moving. The early detection of wheel-flat severity is crucial for passenger comfort and the safety of the railway operation. However, it is still challenging to quantify the properties of wheel flats (e.g., sizes) without interrupting the operations. One way is to transform this damage detection task into a model updating (parameter identification) task. In this abstract, a deep-learning approach is adopted to solve this inverse problem. It has been successfully applied to a field train track test in Singapore. The identified damage size obtained is consistent with on-site measurements.

Files

JIRC2020_B.w3_CAO_FINAL.pdf
(pdf | 0.723 Mb)
License info not available