Jeong Seo Koo

In this paper, research was conducted on anomaly detection of wheel flats. In the railway sector, conducting tests with actual railway vehicles is challenging due to safety concerns for passengers and maintenance issues as it is a public industry. Therefore, dynamics software was utilized. Next, STFT (short-time Fourier transform) was performed to create spectrogram images. In the case of railway vehicles, control, monitoring, and communication are performed through TCMS, but complex analysis and data processing are difficult because there are no devices such as GPUs. Furthermore, there are memory limitations. Therefore, in this paper, the relatively lightweight models LeNet-5, ResNet-20, and MobileNet-V3 were selected for deep learning experiments. At this time, the LeNet-5 and MobileNet-V3 models were modified from the basic architecture. Since railway vehicles are given preventive maintenance, it is difficult to obtain fault data. Therefore, semi-supervised learning was also performed. At this time, the Deep One Class Classification paper was referenced. The evaluation results indicated that the modified LeNet-5 and MobileNet-V3 models achieved approximately 97% and 96% accuracy, respectively. At this point, the LeNet-5 model showed a training time of 12 min faster than the MobileNet-V3 model. In addition, the semi-supervised learning results showed a significant outcome of approximately 94% accuracy when considering the railway maintenance environment. In conclusion, considering the railway vehicle maintenance environment and device specifications, it was inferred that the relatively simple and lightweight LeNet-5 model can be effectively utilized while using small images.

This study proposes a theoretical method to estimate the impact force of Derailment Containment Provisions (DCPs) for the prevention of secondary collisions in the event of a train derailment. By comparing the impact forces estimated using the commonly used Olson model and dynamic simulations, the study identifies significant differences in average and maximum impact forces. The study shows that these differences arise due to the mass effects of vehicle bodies transmitted to the DCP during a collision. To address this issue, the impact force of the Olson model was modified by considering the stiffness of suspensions between masses as a simplified spring–mass model. The modified impact force was verified through impact simulations using the KTX model on curved tracks with various radii. The results show that the modified Olson model provides a reasonable estimate of the impact force, with differences of less than 8% observed under all simulation conditions. This study provides a valuable contribution to the design and analysis methodology for DCPs, improving their effectiveness in preventing secondary collisions and enhancing railway safety.

In this paper, anomaly detection of wheel flats based on signal processing and deep learning techniques is analyzed. Wheel flats mostly affect running stability and ride comfort. Currently, domestic railway companies visually inspect wheel flats one by one with their eyes after railway vehicles enter the railway depots for maintenance. Therefore, CBM (Condition-Based Maintenance) is required for wheel flats resolution. Anomaly detection for wheel flat signals of railway vehicles using Order analysis and STFT (Short Time Fourier Transform) is studied in this paper. In the case of railway vehicles, it is not easy to obtain actual failure data through running vehicles in a university laboratory due to safety and cost issues. Therefore, vibration-induced acceleration was obtained using a multibody dynamics simulation software, SIMPACK. This method is also proved in the other paper by rig tests. In addition, since the noise signal was not included in the simulated vibration, the noise signal obtained from the Seoul Metro Subway Line 7 vehicle was overlapped with the simulated one. Finally, to improve the performance of both detection rate and real-time of characteristics based on existing LeNet-5 architectures, spectrogram images transformed from time domain data were proceeded with the LeNet deep learning model modified with the pooling method and activation function. As a result, it is validated that the method using the spectrogram with a deep learning approach yields higher accuracy than the time domain data.

In this paper, theoretical derailment equations for cross-wind with frequency were derived to assess running safety. For a KTX (Korean high-speed train) unit, the wheel unloading ratios, which are the criteria for evaluating derailments in UIC (International union of railways) and TSI (Technical Specification for Interoperability) regulations, were calculated through the formula under the driving regulations according to cross-wind speeds, and the theoretical results were compared and evaluated through a multibody dynamics (MBD) simulation. In addition, the wheel unloading ratios were calculated for various frequencies of cross-winds. As a result of the formula and MBD, the wheel unloading ratios were shown to increase rapidly regardless of the dampers in suspension when the cross-wind frequency and the natural frequency of a vehicle were in agreement. Finally, we calculated the changes of wheel unloading ratio for different track gauges and found that these theoretical equations could calculate more accurate results than the existing Kunieda’s formula. The formula derived in this study has the advantage of considering various variables, such as fluctuant cross-winds, rail irregularities, and derailment behaviors, which were not considered in previous studies or Kunieda’s formula. It could be used for setting suspensions or railway vehicle specifications in the initial design stage.

This paper studied two useful vibration signal processing methods for detection and diagnosis of wheel flats. First, the cepstrum analysis method combined with order analysis was applied to the vibration signal to detect periodic responses in the spectrum for a rotating body such as a wheel. In the case of railway vehicles, changes in speed occur while driving. Thus, it is difficult to effectively evaluate the flat signal of the wheel because the time cycle of the flat signal changes frequently. Thus, the order analysis was combined with the existing cepstrum analysis method to consider the changes in train speed. The order analysis changes the domain of the vibration signal from time domain to rotating angular domain to consider the train speed change in the cepstrum analysis. Second, the cross correlation analysis method combined with the order analysis was applied to evaluate the flat signal from the vibration signal well containing the severe field noise produced by the vibrations of the rail irregularities and bogie components. Unlike the cepstrum analysis method, it can find out the wheel flat size because the flat signal linearly increases to the wheel flat. Thus, it is more effective when checking the size of the wheel flat. Finally, the data tested in the Korea Railroad Research Institute were used to confirm that the cepstrum analysis and cross correlation analysis methods are appropriate for not only simulation but also test data.

As the operating speed of a train increases, there is a growing interest in reducing damage caused by derailment and collision accidents. Since a collision with the surrounding structure after a derailment accident causes a great damage, protective facilities like a barrier wall or derailment containment provision (DCP) are installed to reduce the damage due to the secondary collision accident. However, the criteria to design a protective facility such as locations and design loads are not clear because of difficulties in predicting post-derailment behaviors. In this paper, we derived a simplified frame model that can predict post derailment behaviors in the design phase of the protective facilities. The proposed vehicle model can simplify for various frames to reduce the computation time. Also, the actual derailment tests were conducted on a real test track to verify the reliability of the model. The simulation results of the proposed model showed reasonable agreement to the test results.

Since the tension clamp is essential for the safe operation of trains, the problem of the broken tension clamp must be carefully managed. The damage of the tension clamp can be caused by various reasons such as track conditions and dynamic behaviours of trains. In this paper, the improvement of durability through the improvement of fatigue strength was studied by considering the experiment results and the FE analyses. First, the analytical model was verified by comparing the results of the stress and strain distributions with the initial clamping force and the results of the FE analyses. And then, the method to improve the fatigue strength characteristics was analytically evaluated by considering the behaviour of tension clamp (middle band and spring arm) and work hardening characteristics.