Jingyuan, Yang ORCID: 0000-0002-7206-800X
(2023).
Inspection of corrosion on the underside of rails based on an acoustic technique.
University of Birmingham.
Ph.D.
![]() |
Yang2023PhD.pdf
Text - Accepted Version Restricted to Repository staff only until 15 March 2025. Available under License All rights reserved. Download (6MB) |
Abstract
The railway is a common global transportation mechanism for people’s daily life, work, and travel. As society has continued to develop, the volume of railway transportation has increased rapidly, making the railway an essential infrastructure in transportation systems. Ensuring the safety of railway operations is crucial for maintaining regular railway services.
Currently, condition monitoring approaches are being rapidly improved to automatically monitor regular railway operations. These approaches include ultrasonic, acoustic, vibration, electromagnetic, laser, and thermographic ones. Automatic railway inspection techniques have the advantage of being able to detect various types of faults using different approaches. However, each technique has its limitations when it comes to identifying corrosion on the underside of the rail effectively.
This PhD thesis focuses on the inspection of corrosion on the underside of the rail using acoustic techniques. To achieve this objective, it introduces methodologies for simulation, feature extraction, and fault classification.
Firstly, inspection of corrosion on the underside of the rail requires sufficient acoustic data for the analysis. However, it is difficult to obtain them from the real world. Hence, a multi-physical Finite Element Model (FEM) is designed to simulate the generation of an acoustic signal caused by the impact of a hammer on the rail. The thesis explains the method for coupling two different simulated physical fields and describes the validation of the simulation model through laboratory tests. A dataset of simulated signals is generated using the validated model.
And then, in order to extract features from the simulated signals obtained from the simulation model and laboratory tests, the thesis compares the effectiveness of different signal decomposition methods for analysis. The most effective decomposition method was enhanced through a proposed method for automatically choosing the appropriate initial parameters. By evaluating the decomposition results, the effectiveness of the proposed method was proved.
Furthermore, the thesis develops a novel integrated classification model to classify the different features of healthy and faulty rails. The results of the classification demonstrate that corrosion on the underside of the rail can be identified through the developed integrated 1-D CNN model based on acoustic impact response signals. Additionally, according to the classification, the simulation model is validated for identifying corrosion on the underside of the rail in real-world scenarios.
This thesis makes three main contributions:
1) The development of an FEM-based simulation model that creates signals that are difficult to obtain in the real world. This simulation model can identify corrosion on the underside of the rail in real-world scenarios.
2) The proposal of a signal decomposition method that effectively extracts features from the impact response signal.
3) The introduction of a classification method that enables the identification of types and levels of corrosion on the underside of the rail simultaneously
Type of Work: | Thesis (Doctorates > Ph.D.) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
Award Type: | Doctorates > Ph.D. | |||||||||
Supervisor(s): |
|
|||||||||
Licence: | All rights reserved | |||||||||
College/Faculty: | Colleges (2008 onwards) > College of Engineering & Physical Sciences | |||||||||
School or Department: | Department of Electronic, Electrical and Systems Engineering | |||||||||
Funders: | None/not applicable | |||||||||
Other Funders: | self funded | |||||||||
Subjects: | T Technology > T Technology (General) T Technology > TA Engineering (General). Civil engineering (General) |
|||||||||
URI: | http://etheses.bham.ac.uk/id/eprint/14381 |
Actions
![]() |
Request a Correction |
![]() |
View Item |
Downloads
Downloads per month over past year