Chi Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Chi Yang has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Korkut Kaynardag
Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin, Austin, USA
7 shared publications - Salvatore Salamone
The University of Texas at Austin, USA
6 shared publications - Guan-Wei Lee
The University of Texas at Austin Department of Civil, Architectural and Environmental Engineering, Smart Structures Research Laboratory 10100 Burnet Road, Building 24, Austin, TX 78758
1 shared publication Junghoon Sohn
The University of Texas at Austin, USA
1 shared publicationKorkut Kaynardag
The University of Texas at Austin, USA
1 shared publicationS. Salamone
Associate Professor, Dept. of Civil, Architectural and Environmental Engineering, Univ. of Texas at Austin, 10100 Burnet Rd., Bldg. 24, Austin, TX 78758
1 shared publication
Latest Publications
- Open Access
- 12 November 2025
- Published by SAGE Publications in Journal of Intelligent Material Systems and Structures
- Vol. 37 (4) , 230-245
Transfer learning in LSTM autoencoders for rail anomaly detection across diverse operational conditions
In this study, transfer learning was applied to a Long Short-Term Memory Autoencoder for the detection of rail anomalies across diverse operational conditions. Separate models were trained initially on Laser Doppler Vibrometer data collected from two distinct sites: the Transportation Technology Center in Pueblo, Colorado, and a BNSF railway in Cleburne, Texas. Each site was characterized by unique environmental setups and train speeds (e.g., 10 mph, 20 mph, 30 mph). A cross-dataset evaluation was performed, testing each model on data excluded from its training. Accuracy was observed to decrease in these evaluations, particularly when speed differences exceeded 10 mph or experimental setups differed. Transfer learning was then implemented, retraining portions of the pre-trained models with the new datasets. The results showed that performance has significantly improved after the implementation of transfer learning. It indicates that transfer learning is essential for ensuring accuracy and adaptability across varying conditions for practical applications.
- 12 November 2025
- Published by SAGE Publications in Journal of Intelligent Material Systems and Structures
- Vol. 37 (4) , 230-245
Open AccessTransfer learning in LSTM autoencoders for rail anomaly detection across diverse operational conditions
In this study, transfer learning was applied to a Long Short-Term Memory Autoencoder for the detection of rail anomalies across diverse operational conditions. Separate models were trained initially on Laser Doppler Vibrometer data collected from two distinct sites: the Transportation Technology Center in Pueblo, Colorado, and a BNSF railway in Cleburne, Texas. Each site was characterized by unique environmental setups and train speeds (e.g., 10 mph, 20 mph, 30 mph). A cross-dataset evaluation was performed, testing each model on data excluded from its training. Accuracy was observed to decrease in these evaluations, particularly when speed differences exceeded 10 mph or experimental setups differed. Transfer learning was then implemented, retraining portions of the pre-trained models with the new datasets. The results showed that performance has significantly improved after the implementation of transfer learning. It indicates that transfer learning is essential for ensuring accuracy and adaptability across varying conditions for practical applications.
- 23 October 2024
- Published by ASME International in Journal of Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems
- Vol. 8 (3) , 1-31
Long Short-Term Memory Autoencoder for Anomaly Detection in Rails Using Laser Doppler Vibrometer Measurements
This study presents an application of a long short-term memory autoencoder (LSTM AE) for the detection of broken rails based on laser Doppler vibrometer (LDV) measurements. This work is part of an ongoing project aimed at developing a noncontact damage detection system using LDV measurements. The damage detection system consists of two LDVs mounted on a moving rail car to measure vibrations induced on the rail head. Field tests were carried out at the Transportation Technology Center (TTC) in Pueblo, CO, to collect the vibrational data. This study focused on the detection of broken rails. To simulate the reflected and transmitted waves induced by the broken rail, a welded joint was used. The data were collected from moving LDV measurements, in which the train was operating at three different speeds: 16 km/h (10 mph), 32 km/h (20 mph), and 48 km/h (30 mph). After obtaining the data, filtering and signal processing were applied to obtain the signal features in time and frequency domains. Next, correlation analysis and principal component analysis were carried out for feature selection and dimension reduction to determine the input used to train and test the LSTM AE model. In this study, the LSTM AE models were trained based on different data sets for anomaly detection. Consequently, an automatic anomaly detection approach for anomaly detection based on the LSTM AE model was evaluated. The results show that the LSTM AE model can efficiently detect the anomaly based on the selected features at three different speeds.
- 23 October 2024
- Published by ASME International in Journal of Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems
- Vol. 8 (3) , 1-31
Long Short-Term Memory Autoencoder for Anomaly Detection in Rails Using Laser Doppler Vibrometer Measurements
This study presents an application of a long short-term memory autoencoder (LSTM AE) for the detection of broken rails based on laser Doppler vibrometer (LDV) measurements. This work is part of an ongoing project aimed at developing a noncontact damage detection system using LDV measurements. The damage detection system consists of two LDVs mounted on a moving rail car to measure vibrations induced on the rail head. Field tests were carried out at the Transportation Technology Center (TTC) in Pueblo, CO, to collect the vibrational data. This study focused on the detection of broken rails. To simulate the reflected and transmitted waves induced by the broken rail, a welded joint was used. The data were collected from moving LDV measurements, in which the train was operating at three different speeds: 16 km/h (10 mph), 32 km/h (20 mph), and 48 km/h (30 mph). After obtaining the data, filtering and signal processing were applied to obtain the signal features in time and frequency domains. Next, correlation analysis and principal component analysis were carried out for feature selection and dimension reduction to determine the input used to train and test the LSTM AE model. In this study, the LSTM AE models were trained based on different data sets for anomaly detection. Consequently, an automatic anomaly detection approach for anomaly detection based on the LSTM AE model was evaluated. The results show that the LSTM AE model can efficiently detect the anomaly based on the selected features at three different speeds.
- 21 November 2023
- Published by Springer Nature
- p. 151-154
Detection of Missing Rail Fasteners Using Train-Induced Ultrasonic Guided Waves: A Numerical Study
- 21 November 2023
- Published by Springer Nature
- p. 151-154
Detection of Missing Rail Fasteners Using Train-Induced Ultrasonic Guided Waves: A Numerical Study
- 15 July 2023
- Published by Springer Nature in Journal of Nondestructive Evaluation
- Vol. 42 (3) , 1-18
Missing Rail Fastener Detection Based on Laser Doppler Vibrometer Measurements
- 15 July 2023
- Published by Springer Nature in Journal of Nondestructive Evaluation
- Vol. 42 (3) , 1-18
Missing Rail Fastener Detection Based on Laser Doppler Vibrometer Measurements
- 21 April 2023
- Published by Elsevier in NDT & E International
- Vol. 137
A rail defect detection system based on laser Doppler vibrometer measurements
- 21 April 2023
- Published by Elsevier in NDT & E International
- Vol. 137
A rail defect detection system based on laser Doppler vibrometer measurements
- 23 January 2023
- Published by Springer Nature in Acta Mechanica
- Vol. 235 (3) , 1453-1469
Investigation of wave propagation and attenuation in periodic supported rails using wave finite element method
- 23 January 2023
- Published by Springer Nature in Acta Mechanica
- Vol. 235 (3) , 1453-1469
Investigation of wave propagation and attenuation in periodic supported rails using wave finite element method
- 1 September 2022
- Published by American Society of Civil Engineers (ASCE) in Journal of Engineering Mechanics
- Vol. 148 (9) , 04022048
Evaluation of Fastening Modeling Approaches for Dynamic Assessment of Rail Based on Finite-Element Method
Fastening systems play a significant role in railroad track dynamics. In this paper, three approaches are investigated for rail fastenings modeling using FEM: (1) spring–damper pairs covering the areas representing the fastening connections; (2) a solid-load approach that explicitly models the solid baseplates and includes toe loads by applying external loads; and (3) a solid-spring approach that takes into account the coupling of rail with the ties by using prestressed springs. Experimental tests were carried out to evaluate the performance of the FEM models. The mode shapes obtained from the FEM models and the experiments were used for the evaluation. A measurement quantifying the similarity of the mode shapes was proposed, which considers both shape similarity and distance similarity. The results showed that the rail foot’s mode shapes obtained from the model using the spring–damper approach demonstrated the most similarity with the ones obtained from the experiments.
- 1 September 2022
- Published by American Society of Civil Engineers (ASCE) in Journal of Engineering Mechanics
- Vol. 148 (9) , 04022048
Evaluation of Fastening Modeling Approaches for Dynamic Assessment of Rail Based on Finite-Element Method
Fastening systems play a significant role in railroad track dynamics. In this paper, three approaches are investigated for rail fastenings modeling using FEM: (1) spring–damper pairs covering the areas representing the fastening connections; (2) a solid-load approach that explicitly models the solid baseplates and includes toe loads by applying external loads; and (3) a solid-spring approach that takes into account the coupling of rail with the ties by using prestressed springs. Experimental tests were carried out to evaluate the performance of the FEM models. The mode shapes obtained from the FEM models and the experiments were used for the evaluation. A measurement quantifying the similarity of the mode shapes was proposed, which considers both shape similarity and distance similarity. The results showed that the rail foot’s mode shapes obtained from the model using the spring–damper approach demonstrated the most similarity with the ones obtained from the experiments.
- 5 June 2022
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2676 (12) , 166-179
Numerical Simulations to Examine the Interaction of Train-Induced Guided Waves With Transverse Cracks
This paper presents the results of a numerical study aimed at investigating the propagation of guided waves generated by train wheels on rails. The particular application of interest is to exploit such waves for the identification of transverse cracks located in the rail head. The research presented is part of an ongoing project aimed at developing a non-contact damage detection system based on laser Doppler vibrometer measurements. In this study, numerical simulations were carried out using three rail models. The first two models consisted of a transverse crack whose size is 20% and 10% of the rail head cross-section (RHC), respectively. The third model consisted of the same crack size as the first model, but it was designed to be longer before and shorter after the location of the crack compared with the first model. The goal of this model was to examine the location where the effect of cracks on the waves first appeared. To acquire train-induced guided waves to use in the simulations, an accelerometer was placed under a rail head, and propagating waves were recorded during the passage of an operating train. Afterward, on each rail model, the location of excitation and two measurement points were shifted forward over the rail to replicate the movement of the defect detection system. Two damage functions were used to examine the change in wave propagation caused by cracks, while a multi-dimensional damage index consisting of the damage functions was used to identify the location of the crack (20% of RHC).
- 5 June 2022
- Published by SAGE Publications in Transportation Research Record: Journal of the Transportation Research Board
- Vol. 2676 (12) , 166-179
Numerical Simulations to Examine the Interaction of Train-Induced Guided Waves With Transverse Cracks
This paper presents the results of a numerical study aimed at investigating the propagation of guided waves generated by train wheels on rails. The particular application of interest is to exploit such waves for the identification of transverse cracks located in the rail head. The research presented is part of an ongoing project aimed at developing a non-contact damage detection system based on laser Doppler vibrometer measurements. In this study, numerical simulations were carried out using three rail models. The first two models consisted of a transverse crack whose size is 20% and 10% of the rail head cross-section (RHC), respectively. The third model consisted of the same crack size as the first model, but it was designed to be longer before and shorter after the location of the crack compared with the first model. The goal of this model was to examine the location where the effect of cracks on the waves first appeared. To acquire train-induced guided waves to use in the simulations, an accelerometer was placed under a rail head, and propagating waves were recorded during the passage of an operating train. Afterward, on each rail model, the location of excitation and two measurement points were shifted forward over the rail to replicate the movement of the defect detection system. Two damage functions were used to examine the change in wave propagation caused by cracks, while a multi-dimensional damage index consisting of the damage functions was used to identify the location of the crack (20% of RHC).