Rail track condition monitoring: a review on deep learning approaches (original) (raw)

A Review on Deep Learning Techniques for Railway Infrastructure Monitoring

IEEE Access

In the last decade, thanks to a widespread diffusion of powerful computing machines, artificial intelligence has been attracting the attention of the academic and industrial worlds. This review aims to understand how the scientific community is approaching the use of deep-learning techniques in a particular industrial sector, the railway. This work is an in-depth analysis related to the last years of the way this new technology can try to provide answers even in a field where the primary requirement is to improve the already very high levels of safety. A strategic and constantly evolving field such as the railway sector could not remain extraneous to the use of this new and promising technology. Deep learning algorithms devoted to the classification, segmentation, and detection of the faults that affect the railway area and the overhead contact system are discussed. The railway sector offers many aspects that can be investigated with these techniques. This work aims to expose the possible applications of deep learning in the railway sector established on the type of recovered information and the type of algorithms to be used accordingly.

Improving Railway Safety with AI: Detecting Defective Fasteners using Deep Learning

Zenodo (CERN European Organization for Nuclear Research), 2023

Railway track inspection is a vital task to ensure safe and efficient train travel. However, traditional inspection methods require people to walk along the tracks and look for issues, which is time-consuming and exhausting. To address this, we created a new system that uses a camera installed under a train to take pictures of the tracks automatically. Then, we applied advanced computer methods like YOLO (You Only Look Once), SSD (Single Shot MultiBox Detector), and ResNet50 (Residual Network50) to analyze the images and identify any areas that need fixing. ResNet50 is a type of computer network that can recognize patterns in pictures, and we used it to enhance the performance of the other two methods. Our system achieved an average detection rate of 99%, meaning that it can identify track anomalies as accurately as a human operator but much faster. This new system has several benefits over traditional methods, such as increasing efficiency and accuracy, minimizing human involvement, and reducing operational costs. By improving safety and reducing the risk of accidents, this technology has the potential to revolutionize the railway maintenance industry. In summary, our study presents a new automated technique for detecting rail anomalies using computer vision and deep learning-based methods. This technology can help make train travel safer and more efficient, ultimately benefiting both passengers and railway companies.

Developing Machine Learning-Based Models for Railway Inspection

Applied Sciences

Smart railway maintenance is crucial to the safety and efficiency of railway operations. Successful deployment of technologies such as condition-based monitoring and predictive maintenance will enable railway companies to conduct proactive maintenance before defects and failures take place to improve operation safety and efficiency. In this paper, we first propose to develop a classification-based method to detect rail defects such as localized surface collapse, rail end batter, or rail components—such as joints, turning points, crossings, etc.—by using acceleration data. In order to improve the performance of the classification-based models and enhance their applicability in practice, we further propose a deep learning-based approach for the detection of rail joints or defects by deploying convolutional neural networks (CNN). CNN-based models can work directly with raw data to reduce the heavy preprocessing of feature engineering and directly detect joints located on either the lef...

Railway Track Inspection Using Deep Learning Based on Audio to Spectrogram Conversion: An on-the-Fly Approach

Sensors

The periodic inspection of railroad tracks is very important to find structural and geometrical problems that lead to railway accidents. Currently, in Pakistan, rail tracks are inspected by an acoustic-based manual system that requires a railway engineer as a domain expert to differentiate between different rail tracks’ faults, which is cumbersome, laborious, and error-prone. This study proposes the use of traditional acoustic-based systems with deep learning models to increase performance and reduce train accidents. Two convolutional neural networks (CNN) models, convolutional 1D and convolutional 2D, and one recurrent neural network (RNN) model, a long short-term memory (LSTM) model, are used in this regard. Initially, three types of faults are considered, including superelevation, wheel burnt, and normal tracks. Contrary to traditional acoustic-based systems where the spectrogram dataset is generated before the model training, the proposed approach uses on-the-fly feature extract...

Multi-Segment Deep Convolution Neural Networks to Classify Fault in Sensor Reading of Railway Point Systems

2019

Intelligent fault detection by sensor data can ensure the reliability and availability of critical infrastructures. Switches and crossings (S&C) are one of the most important assets of railway networks. They divert trains in different directions by shifting the position of switch rail by point operating equipment (POE). The sensors record the electrical current drawn by the motor in POE. The extraction of features from time-series sensor data enables the detection of faults in POE. This paper proposes a deep learning model to detect faults in railway POE without the need for preprocessing the raw time-series data. It is based on 1-D convolution neural network. The architecture of the proposed deep learning network consists of three types of layers. The first layer is called the local convolution layer. It consists of three 1-D convolution layer to extract local temporal features from three non-overlapping segments of time-series data of different operating phases of POE. The second layer is fully-connected convolution layer. It extracts global temporal features. And the last layer is the output layer, it provides the binary output of fault or fault free for a given sensor data. The result shows that this framework can classify fault with 95.60% accuracy.

Multi-Segment Deep Convolution Neural Networks for Classification of Faults in Sensors at Railway Point Systems

25th International Conference on Automation & Computing, Lancaster University, 2019

Application of Deep Learning Networks to Segmentation of Surface of Railway Tracks

Sensors

The article presents a vision system for detecting elements of railway track. Four types of fasteners, wooden and concrete sleepers, rails, and turnouts can be recognized by our system. In addition, it is possible to determine the degree of sleeper ballast coverage. Our system is also able to work when the track is moderately covered by snow. We used a Fully Convolutional Neural Network with 8 times upsampling (FCN-8) to detect railway track elements. In order to speed up training and improve performance of the model, a pre-trained deep convolutional neural network developed by Oxford’s Visual Geometry Group (VGG16) was used as a framework for our system. We also verified the invariance of our system to changes in brightness. To do this, we artificially varied the brightness of images. We performed two types of tests. In the first test, we changed the brightness by a constant value for the whole analyzed image. In the second test, we changed the brightness according to a predefined ...

Fault Detection from Images of Railroad Lines Using the Deep Learning Model Built with the Tensorflow Library

Turkish Journal of Science and Technology, 2022

A means of transportation is the way in which an object, person, or service is transported from one place to another. Rail transportation occupies an important place in terms of cost and reliability. Most train accidents are caused by faults in railroad tracks. Detecting faults in railroad tracks is a difficult and time-consuming process compared to conventional methods. In this study, an artificial intelligence based model is proposed that can detect faults in railroad tracks. The dataset used in the study consists of defective and non-defective railroad images. The proposed model consists of foldable neural networks developed using the Tensorflow library. Softmax method was used as a classifier. An overall accuracy of 92.21% was achieved in the experiment.

Deep learning-based fault detection in railway wheelsets using time series analysis

Mehran University Research Journal of Engineering and Technology

Maintenance of Railway rolling stock is usually scheduled based. However, the mechanical parts, especially the wheelset may wear down prematurely due to several factors such as excessive braking and traction forces and environmental conditions. This makes the scheduled maintenance less effective and sometimes it results in derailments. This paper presents a deep learning-based technique to detect wheel conditions so that maintenance can be performed promptly and efficiently. A time series dataset of axle vibrations is generated using a simulation model of the wheelset. The dataset is then used to train and test the deep learning model. Long short-term memory (LSTM) architecture is selected for this application since it is designed to perform better for time series datasets. The results show good performance in terms of training and testing accuracy. The model is tested in different defect scenarios and the mean square error in the prediction of railway wheelset parameters is around ...

Rail track condition monitoring: a review on deep learning approaches (original) (raw)

Related papers