Good vibrations – Distributed acoustic sensing for train tracksProfiles
The Device Chronicle interviewed Sebastian Haid, Product owner, Team Train Tracking, Sensonic to learn about how distributed acoustic sensing is advancing the sensing of potential adverse impacts on rail tracks.
Sebastian trained as an astrophysicist and graduated with expertise in magneto-hydrodynamics, big data analysis, high-performance computing, simulations and theoretical physics. Sensonic is part of the Frauscher group and develops railway specific solutions based on fibre optic sensing. FOS is the best technology for sensing for enabling accurate quantitative measurement and analysis of real rail events such as track vibration.
Distributed acoustic sensing comes to rail
Sebastian begins by explaining that train tracking information provisioning is a growing field. Distributed Acoustic Sensing and fibre optic sensing has been used in many industries since the 1970s for application in oil & gas such as pipeline monitoring and perimeter defense. Its use in rail came later in 2014 when Frauscher – the parent company of Sensonic introduced a sensor product into the segment. Sensonic has since innovated further and by collaborating with 12 progressive partners from the sector, Sebastian and his product and operations team have adapted the sensor technology to make appropriate solutions for important use cases such as condition monitoring. The partners provide access to the base information and then Sensonic performs the data processing and analysis. The partners include freight lines with single tracks and 2km long trains full of iron ore, regional lines carrying passengers; mainlines with mixed traffic and finally high speed lines with high speed passenger trains. Sebastian explains that Sensonic addresses all the rail segments with the exception of metros and tramways. He explains the challenge here: “Laying fibre to do sensing on trams is very difficult as these vehicles share the same road space as heavy trucks and so it is very challenging to disambiguate vibrations coming from the two different sources in this case.”
Explaining how distributed acoustic sensing works
Next, Sebastian explains how acoustic sensing works. It is, he says, a “80km long microphone which is one sensor. Whenever or wherever something is happening on or near the track, we can hear or receive the vibrations, we get less than 10 terabytes of data per day and at the sensor box, we must extract insights from the data. The Sensonic team identifies trains and the wheel-rail interaction. Then from the data, insights can be gathered to answer a series of critical questions such as:
- Is there something wrong with the wheel or the rail track?
- Are there people trespassing over the track
- Are their people working near the track?
- Has there been a rock fall near the track?
- Was there a short circuit between the pantograph and the electricity line?
Sebastian and his team track all of this information which is pre-processed with AI. He explains “The train will run from A to B so generally so the data processing for train tracking in this regard is not particularly challenging. It gets more challenging when you look at the installation itself. We apply a digital twin where we take the train as a sensor and run it over the installation, calculate some metrics and take the information from below the train. We then compare that data across 1000 trains that have passed over the installation, then you can see change. You are looking for sudden change, such as a dramatic and impactful change that has happened when the last train passed over the installation in comparison to the train that passed over it previously. If there has been a sudden change, this could mean that a rail track break has occurred.” Sebastian also says that in advanced rail markets such as central Europe, there is a misnomer that rail breaks do not occur. But in fact in most cases where rail breaks do occur, they are identified quickly enough and solved and repaired before something dramatic happens. In other regions of the world this is not the case and a fatal accident occurs. The classic way to find rail track breaks is for an operator to send employees along the track twice a week to perform manual checks on the tracks. A more advanced way to do it is to have a track circuit, connecting electricity into the rails, one left and one right and when the train runs over it, if there is a break, there will be no current. Sebastian explains that the drawback with this approach is that rain or fallen leaves can interfere with the accuracy. “On the other hand, Glass fibre doesn’t care about magnetic fields, or rain or stones falling on it” he concludes.
Acquiring and processing the data
Acoustic sensing is the leading edge approach but 8.5 terabytes of data is collected from each sensor and sensors can run side by side. The data amount can be reduced by only focusing on the train-related information, and this data can be stored in the cloud. For the machine learning models, label data is required so this data will be transferred to the cloud as the model training takes time and cannot be done on premise. Sebastian explains that the cloud provides the resources his product team needs. The challenge is that the rail operators have some reservations about the cloud. Sebastian says the operators can “think in terms of boxes carrying data packets. This is where the partners are important as they are willing to take the step and connect a sensor to the cloud. Many operators are more conservative, risk averse and are afraid that a third party could enter a sensor and make modifications. But the change is happening as more executives see the benefits of digital adoption.”
Updating the sensor boxes
Currently, Sebastian explains that a VPN connection is used to apply changes to the Sensonic system. OTA updates are at an early stage at Sensonic. Sebastian explains that currently, AWS update manager is used, a model is trained and then it is deployed to a repository which is then brought automatically to the box in the field. Sebastian explains a challenge that the large rail operators just want to provide one time access at certain times whereas Sensonic wants to deploy as many times and as quickly as possible to provide the best information services possible.” Sebastian says Sensonic has been thinking about over the air updating of the remote devices. Sebastian says “If you take a remote position, how do you get information from there and to there? We have to get information to the sensors as the scale of the fleet grows. From a product management perspective, every single update should be done remotely, we do not want to send any engineers on site, as this leads to a significant cost. The Pandemic has taught us that boxes can be deployed remotely and that it works. The update of the FPGA is challenging. The FPGA calculates from the optical signal for the first pre-processing step. You need this to extract information from the optical sensors.”
We wish Sebastian and his colleagues well as they advance the use of acoustic sensing and AI to serve more use cases in train and track management.