Digital Twin of German Railways Based on Fiber Optic Sensing Technology

Deutsche Bahn AG(DB) is currently studying the application of optical fiber sensing, digital twins, and other technologies to improve the transportation capacity of the German railway network and provide support for train mobile block operation and intelligent maintenance of railway infrastructure. The article will introduce the FOSSIL4.0 research project jointly carried out by DB Netz and its partners. The project aims to explore the potential uses and functions of fiber optic sensing(FOS) technology to replace traditional passive sensors to provide the required data and create a new, more accurate, and sensitive nervous system for railway digital twins.

1. Background

A digital twin is a simulation process that makes full use of physical models, sensor update data, and operating history data, integrating multi-disciplinary, multi-physical quantities, multi-scale, and multi-probability. It can complete the mapping in the virtual space, thereby reflecting the entire life of the corresponding physical equipment. In a cyclical process, almost any object in the real world can create a digital twin. Its essence is that computer-based models receive data and information through the latest communication technologies, apply analytics software to perform real-time optimization based on this, and then transmit the output back to the real world. The application of digital twins has become an inevitable trend to promote “industry4.0”.

With the maturity and promotion of related technologies, digital twins are gradually applied in the railway field. Railway infrastructure companies and operating companies have placed great expectations on using digital twins to obtain detailed real-time information about the railway system, improve the transportation capacity of the railway network, shorten the maintenance time of the railway system, and achieve predictive maintenance.

Building a railway digital twin requires the support of a large amount of real-time data. The more sensitive and comprehensive the sensor network responsible for collecting data, and the more timely the data collected, the better it can map reality. However, this is a [particular challenge for DB. DB’s line network subsidiary (DB Netz) operates one of the largest railway lines in Europe, with a total line mileage of more than 33,000 kilometers. Installing power and communication links for each passive sensor in such a large-scale wire network requires enormous manpower, material, and financial resources. DB Netz, therefore, sought to find suitable alternative sensor technologies to support its digitalization strategy. Currently, it is cooperating with AP Sensing company and the Darmstadt University of Technology to carry out the FODDIL4.0 research project, aiming to explore the potential uses and functions of FOS technology to replace traditional passive sensors and provide the required data.

2.FOSSIL4.0 project overview

The FOSSIL4.0 project was launched in April 2020 and is expected to run until March 2023 with a budget of 3 million euros, two-thirds of which come from the German Federal Ministry of Digitalization and Transport(BMDV). The main research content of this project is to explore how to use fiber optic sensing technology to support various railway applications, which will help to determine whether DB’s existing fiber optic infrastructure can be used for train integrity detection so that trains can run safely under the condition that the running interval is shortened; a detailed collection of real-time status data of trains t support intelligent and predictive maintenance of trains, thereby improving their availability; testing technologies such as energy-saving driving, track obstacle detection, and short-circuit rapid positioning; developing software security architecture and intelligent maintenance, move occlusion algorithms, intelligence(AI) methods, and use AI for automatic security verification.

3. FOS technology

The FOS technology of this project is mainly provided by AP Sensing. This technology uses the principle that the transmission characteristics(such as phase, intensity, etc.) of light in the optical fiber will change with the external environmental factors(such as temperature, pressure, electric field, magnetic field, etc.) of the optical fiber, using an existing fiber optic cable network(currently only used for data transmission) as passive sensors, making it a long chain of thousands of sensors. Just like humans use their senses to perceive the environment and transmit signals to the brain, digital twins can use the FOS system as their own nervous system to collect and record temperature and vibration data at each point along the way detected through optical fibers. The positioning accuracy of the FOS system for various changes can reach the meter level.

At present, optical fiber cables have been laid next to DB’s main railway lines for railway communication and will cover the entire railway network in the next few years. These cables act as FOS devices, collecting all the sound and frequency patterns produced by the train as it travels along the track, providing a temporally and spatially continuous acoustic image of the railway infrastructure and the train. The associated FOS signals, which are always linked to the train’s real-time position, speed and speed, and length, enable the detection and differentiation of multiple trains traveling simultaneously along one track, as well as individual trains traveling on individual tracks of a multi-track line. Since the FOS system can provide accurate position information and status data of all trains on the track, it can be used to comprehensively detect the integrity of the trains, so as to achieve the goal of shortening the interval between trains and improving the transportation capacity of the railway network.

4. Improvement of FOS system sensitivity

Due to the significant improvement and expansion of the sensitivity and coverage of the FOS system, it is now possible to detect small changes in the acoustic signal caused by form trains and locate their source. On this basis, the special adjustment algorithm based on its learning can realize the decoupling of the signal generated by the wheel and the rail track, so as to determine whether the change starts from the train or is caused by the track structure.

In order to prove the sensitivity and effectiveness of the digital twin based on FOS technology, the project personnel conducted relevant tests on a section of the Berlin-Dresden main line north of Barus/Mark. The test results are shown in figure1. The figure shows a series of different acoustic events detected by the FOS system, in which the acoustic signal of simulated fiber optic cable theft is significantly different from the acoustic signal of a train passing by and a cat crossing a level crossing. At 12:27:30, a train pulled into the Dahlewitz station, slowed down, and stopped, and the relevant acoustic signal was clearly visible in area B of the figure; at the same time, a simulated cable theft occurred in an area 1km away from Rangsdorf, see C in the area marked by the red box; 1km away from this is the bridge across the railway on the A10 expressway. Frequent vehicles on the road will affect the acoustic measurement signal, see the large red area in area D. All three events occurred simultaneously at different locations along the cable, and by combining FOS technology with a special adjustment algorithm based on machine learning, it was possible to identify and correctly distinguish the three.

Fiber Optic Sensing Technology

Fiber Optic Sensing Technology

Fig.1 Display of three types of event acoustic signals measured by the FOS system

5. Realize intelligent maintenance of railway infrastructure

One of the advantages of building a railway digital twin is that the detection data of the FOS system can be used for railway infrastructure diagnosis. A railway digital twin can use its “acoustic memory” to analyze the history of the acoustic signal development of one section of the lien and compare it with that of another section to reliably detect slow, wear-related track structure variety. This “acoustic memory” helps enable intelligent maintenance of railway infrastructure by improving rail track availability and traffic capacity.

Although the currently widely used moving block technology can shorten the distance between trains and increase the wear and tear on the railway infrastructure. The railway digital twin based on FOS technology can realize the influence of railway infrastructure maintenance and update the availability of infrastructure, give full play to the advantages of mobile blocking technology, and effectively improve the transportation capacity of the railway network.