Keeping trains on track: How AXO Track prevents railway delays with IoT sensors

Oliver: Yes, exactly. It's where the train changes direction. And the switches are one of the most critical infrastructure parts, and it's a part which wears and tears over time.

Ruth: Okay.

Oliver: You know, the wheel sets of a train jump basically from one rail to the next. And this has enormous forces. This has enormous impacts. So, it's all about the perfect geometry of the switch. So, the transition from one wheel set from one track to the other is very smooth.

Ruth: Okay. Yeah, I can just remember it from model trains. Basically, the little trains derailed always when you tried to switch tracks too fast and then everything flew out in one corner. Right. So, I imagine there's lots of weight, there's lots of force, and there's lots of friction going on. So, is that why they're so sensitive?

Oliver: Yes. Of course, the faster the train, the more forces on it, or the heavier the train. And basically, the whole track system brings vibration into the ground. So, you have train steel on steel, this vibrates, and the vibration goes through the sleepers, through the gravel into the ground. And this gravel starts to - it's really sharp gravel, and it starts to iterate. Then the whole system starts to move, and this gets stronger and stronger over time. And you have to detect when there's a point of basically no return and do maintenance before this happens.

Ruth: And in the course, you took, what did you learn? Is that where you thought about what company you want to build and what product you want to make?

Oliver: Yes, exactly. I really was curious and was wondering why are you not measuring this? There are some parts which are getting measured, but other parts which nobody cares about. This is where I said, okay, let's try, let's have a look. I discussed this problem with specialists at Deutsche Bahn and said, this is something you have to develop on the track. And when I got a pilot track to develop, this was the moment where I said, okay, it makes sense. I'm founding AXO Track.

Ruth: Fascinating. Before we dive into your sensor technology and what it is that you do - had no one ever thought about this before?

Oliver: No. There are solutions in the market to measure it, but not on all critical parts. And this is basically due to development of sensor technology. You always, in a way, could measure it, but to bring out a sensor system which is permanently on the track and can resist these very strong demands - heat, cold, the forces which the train brings in - is difficult and would have been very expensive. And only recently through the development of the smartphone or through the smartphone industry and the development of sensor technology, the technology is price-wise cheap enough that you can employ technologies permanently on the track.

Ruth: Okay. What do you mean by permanently?

Oliver: It's always there. Operating in the best way without having any power source, driven by batteries or solar cells. Solar cells are very difficult on the track because of the dirt. Then of course, you don't want to change batteries every six months, so it should run for years because you cannot just go on the track. You know, if there's an ICE coming by with over 160 kilometres per hour, you don't want to be anywhere near.

Ruth: Yeah, I imagine. So how does your IoT sensor technology work? What parameters are you monitoring on these switches?

Oliver: We are monitoring mainly vibrations, but also temperature and humidity.

Ruth: And this helps you then analyse the health of the switch.

Oliver: Exactly. You can really see every wheel set. We can even identify through the patterns exactly which train is going over the track and at which speed and so on. And there you can really see what the impact on the track is and on the switch.

Ruth: Interesting. I suppose you can also then loop that back and make deductions about what maybe trains need to look like in the future to help sustain critical infrastructure, or is that too far?

Oliver: I think that the rolling stock - as they say, the trains - the railway industry is basically divided in rolling stock and infrastructure, and we are more on the infrastructure side. The rolling stock side already for many years has had a lot of IoT systems in place, and the infrastructure side lags a little bit behind.

Ruth: Okay. Tell us about the technical challenges of deploying the sensors. You just mentioned dust and vibrations. Railway infrastructure operates in extreme conditions and especially in Europe where you have different seasons, I suppose in southern countries it's more heat-related problems. In northern countries it's more frost-related problems. How do you ensure reliable data collection?

Oliver: The data connection is ensured by 5G network, cellular network, and so we send the data over the 5G network. So, if one train passes the switch, you have hundreds of thousands of data points. And it makes no sense to send hundreds of thousands of data points, for example, to Deutsche Bahn. What are they doing with this? So, you have to have an interpretation algorithm. And to keep the battery consumption very low, you interpret this data already on the embedded system and then just send the results.

Ruth: Oh, interesting. So, the system collects the raw sensor data and then turns it into actionable insights for the maintenance teams already on the track.

Oliver: Yes, we can basically ask a system either to send raw data or send results. But in permanent use, they're just sending results of course.

Ruth: And to circle back, you mentioned that smartphone technology or the evolution of smartphone technology helped accelerate your innovation. What other factors have made IoT-based railway monitoring possible now?

Oliver: There's no other way than to look into digitalisation of the railway network. You see, Deutsche Bahn wants to double passengers until 2030. And this is also the political will. You have the Green Deal, which says we want to reduce CO2 emissions by 30% by 2050. And if you want to do this in the transport sector, you need to shift traffic from the plane to the rail, from the road to the rail, and you cannot just double the infrastructure. We have good infrastructure, but this means you need more trains on the tracks at a higher frequency and perhaps also with higher speed. And this increases the stress on the infrastructure, which is, as probably everybody knows, already old.

So, at the same time, you have a shortage of skilled workers. So, you have very experienced, very motivated workers, but a lot of them are going to retire and you don't get enough young people. If you increase the stress on the tracks, you also need to maintain this more. To maintain the track, you need pauses where the trains can't go. If you increase the frequency of the trains, you have fewer pauses to go. So, there is a strong need to go into digitalisation and do this.

At the same time, it's a very regulated market because in the infrastructure, it's all about safety. You want to avoid accidents, so this is why you have rules where track workers go out at certain intervals and go step by step, which is also right to do. And the rate of accidents in Germany is very, very low. So, this is always the first priority - avoid accidents.

So, you have to shift these rules, and this takes time of course to shift rules because you have long trial phases where you see, okay, we are sure that in certain ways sensor technology can take over the view of a railway technician or even improve it.

Ruth: And the issues you mentioned - older infrastructure, lack of investments, I suppose, shortages of skilled workers - that of course Germany is a good example. Obviously since this is where you started the company, but this goes, I think, for most of the European system, right, for France as well, for the UK. I always think that in Japan, the trains are supposed to be really modern and fast and digitalised. Can you tell us a little bit about the market in Europe, the inspiration you're getting from other areas in the world?

Oliver: So, in Japan they have different trains. They also have different materials on the switch and crossing. So, it's slightly different problems. But they still have this problem. So, we are also in discussion with Japanese switch manufacturers. It is actually quite interesting. So, the railway infrastructure - the average age of switches in Germany is over 20 years. And this can be fine if it's taken care of well.

So, the rail equipment is a very long-term planning process, so you need to plan for 10 to 30 years. Very often the political framework doesn't give this long planning period and is not giving the necessary budgets. Or more importantly, the supplier industry doesn't have the capacity. So, the best thing you could do would be just to close the whole network down completely, renovate it for two years and then it's there. But even if you would do it, and even if you would put the funding in, the industry doesn't have the capacity to do it. You have to do this step by step.

And what politics could do is create something like an infrastructure fund like you have in Switzerland, which works very well also in Austria. So, we have a long-term planning horizon. In Germany, you have only 12 months, 12 months, 12 months. And then you have political change. In the supplier industry, they're not increasing production capacities if they don't have the security that there's continuous investment going on.

But actually there's - they're starting, which you might have heard, by renovating these corridors which are the most critical corridors, starting with the so-called line between Frankfurt and Mannheim. And from August on, it'll be the track Berlin-Hamburg, and then there are 40 more high-speed corridors or critical corridors which will get renovated. So, they're working on the problem, but it's nothing which can be cured in six months. This is a project over years.

Ruth: Yeah. And the shortage of skilled workers then obviously does not help at all. But it seems that your technology can help operators deal with this challenge.

Oliver: Yeah, it can make the whole maintenance process more efficient. In Germany, you have 2.5 million minutes per year where switches are not working. So, these are almost five years accumulated.

Ruth: Wow. In one year?

Oliver: Well, what happens if there's a malfunction? The whole track gets closed, and the trains are getting rerouted, and then people have to go out and check the switch and also repair it. And if you don't have enough people, it takes time till they're out on the track and then they have to check it, and then they might change some component, and it might not be in the warehouse. It needs to be ordered. So, this might take time. And of course, the fewer people you have to go out...

So, what you want to avoid is a breakdown. So, what you want to do is you want to know exactly at every minute what is the condition of your switch. And if you see there might be a breakdown, you want to send out service teams in advance. And there might be some switches which are very stable and work very well, and you say, okay, I see this switch is fine. I don't have to send a team at regular intervals to this switch. I'd rather take those resources and already put them on a switch where I see this might have problems in the next three months.

Ruth: And that is the benefit of doing predictive maintenance, right? You invest a little bit, but then in the long run you prevent really critical failures where once the thing breaks, it would shut down the track for a longer period of time, causing even more disturbances.

Oliver: Yes. And also of course, if you start calculating all the delay minutes of the passengers in the train and the economic damage. Of course, it's worth a lot to do research here.

Ruth: We will take a short break, stay with us. We will be hearing from our guest very shortly. This podcast is brought to you by Avnet Silica, the engineers of evolution. We help you bring secure, intelligent, and connected products to market. If you want to learn more about us, we have put information and links in this episode's show notes, and you can also connect with us on LinkedIn or avnet-silica.com. That's AVNET dash SILICA.com.

How do you prove the effectiveness of your system to potential customers? What metrics matter most to railway operators?

Oliver: So, at the moment we have a scientific project together with Fraunhofer IVI and the Federal Railway Agency, Eisenbahn-Bundesamt. And of course, as you can imagine, it takes some time, it's a three-year project to prove what kind of digital systems really work to make it efficient. And at the end, the goal is to reduce this 20% of delays because of switch failures significantly. You will never reduce it 100% because there are failures you cannot foresee, but you can reduce a significant number.

Ruth: Interesting. Is it the Fraunhofer IML for the logistics department?

Oliver: It's IVI. There's deep knowledge about the tracks, and it really makes sense. So, everybody has to work together here, you know, even the different suppliers of sensor technology. There should be an exchange and then you should see what the best solution for which component is.

Ruth: Yeah. It's really interesting that you have the operators, the suppliers, researchers, and the government all pulling together to make the system better in general.

Oliver: Yeah. In Germany, you say "gründlich" - you really have to drill deep down. It's long term, but it's also very rewarding because you can really make an effect. You can really improve the railway system, which I think is on the one hand an old system, but it's a system which is very important going into the future.

Ruth: Yeah. As you mentioned, it is the most sustainable and environmentally friendly way to move a lot of people at the same time. Right. There's nothing compared to CO2 emissions. Trains are just the most efficient and with the lowest CO2 emissions.

Oliver: Yes, if it is electrified. But it's the most energy-efficient way of mass transport. Yeah. Because you have just steel on steel and you hardly have any friction. So, if you imagine pushing a car which weighs 1.5 tonnes, you cannot push it with one finger, but if you have a wheel set of a train which is just on the rails, you can push it with one finger, and it starts to roll.

Ruth: Really?

Oliver: You have to find the wheel set and a track where you can push it.

Ruth: I have to try that next time I'm at the station.

Oliver: You have to find the wheel set. Yeah. And a track where you can push it.

Ruth: Regarding the investment, you are also looking to Austria, to Switzerland, where they have different approaches to infrastructure investments. What lessons are you applying to your European expansion strategy?

Oliver: There's a very tight exchange between Germany, Switzerland, Austria, between the professionals and technicians, but also with France. So, there are certain conferences where there's exchange and we are talking with all of these people and of course they're looking very carefully at what we are doing in Germany.

Ruth: Can you talk a little bit about your future plans?

Oliver: So, our home market is Germany. We really want to improve it. We want to reduce the Weichenstörung, the breakdown of switches and crossings. We want to prove it and then we look at other European markets, which is UK. We already have discussions with France. We are in discussions with Austria and Switzerland. By the way, the railway networks work much better there, but they also have other circumstances, but still, they have the same problems, and you can still improve it.

You know, also in the US the freight lines are very interesting. Also, passenger traffic in the US is growing. The Asian countries are also interesting markets. China's a very interesting market.

Ruth: What role does artificial intelligence play in your predictive maintenance algorithms?

Oliver: I think it's a very important role. It's getting played and it gets more and more important as the technology improves, but you have to make sure that you really understand what is going on the track so you can train the right models. And it's very complex because if you see a car - a car sees a stop sign, and the stop sign is always the same, but on the track, on a switch, no switch is identical to another switch. There's a difference in age, there's a difference in radius, there's a difference in underground, there's more heat on one, more snow on the other. Everything is different. So, you really have to identify the key parameters. And it's not only that you see the moment when the switch breaks down, you really want to come from condition-based maintenance to predictive maintenance as early as possible.

Ruth: And where do you see railway IoT technology headed? What developments are you most excited about?

Oliver: Both sides. You have developments in the infrastructure, and you have developments on the rolling stock and both sides have very interesting developments. So, for example, rolling stock. Do you know at what speed two cargo trains hit each other? Nobody knows it, but if it's too strong, there might be damage. Or what is a solved problem in rolling stock is really monitoring the doors that everything is closed. In infrastructure, there's a lot to do. For example, the catenary - the electricity lines, if there's a short circuit somewhere because a branch falls on it or whatever. The track workers still have to go along the catenary and look where did this happen? And you want to know exactly where's the spot so you can send the teams, and they can repair it.

Another problem is also the network. I mean, you know, it's hard. You have a radio cell, and the train comes with 500 people or more in, and it goes for seconds through the cell. And then it goes to the next cell. And everybody wants to have high-speed internet. You want to continue working with your cloud and it should all work seamlessly. And of course, also for our technology, the network is very important, so there's still a lot of challenge to be done.

Ruth: Yeah, it'll be exciting to see where the digitalisation efforts in this segment will lead in the next years.

Oliver: Yeah. I mean, there are studies which say it can increase rail traffic by 30% - digitalisation of the different systems without increasing the network.

Ruth: Okay. Wow. I would've thought that the infrastructure wouldn't be able to handle that at the moment. But you say that could be done.

Oliver: Yes. It has to be done. It has to be done. You also want to add infrastructure, but it's more an organisational issue and it's also getting the infrastructure renovated and it's like a roof. Our roof leaks at the moment. So now it's worse than keeping the roof always in the best condition. This might be more expensive in the short term, but it's much, much cheaper in the long term for sure. And this is what we're seeing now. We just have to catch up.

Ruth: Absolutely. What's next for AXO Track? Any exciting new developments or partnerships you are allowed to share with us?

Oliver: We are just in development mode. So, we are in a very comfortable, quiet situation, sitting in our office. We are very busy until end of September to bring sensors on the track. As you can imagine, it's not so easy. It's one point - we need to get installed very quickly. You don't want to disturb the train traffic. You just go out, mount the sensor system and then go off the track. And so, this all has to be organised. So, we move forward step by step now. No exciting announcement at the moment.

Ruth: It is new technology, and you are rolling it out. So that is terrific. And then the last and most difficult question, what song would you put on the soundtrack of this episode?

Oliver: This would be "Railroad Man" by Eels.

Ruth: Very nice. Thank you. I will put it on, we have a playlist on YouTube and on Spotify, and it is growing. Great. It is. I love this song. A very nice addition to the compilation. Thank you so much.

Oliver: Thank you.

Ruth: Oliver, it has been a pleasure having you on the show. Thank you so much for sharing your insights. I'm really excited to see where your company and your technology is going. And I wish you all the best of luck.

It was really interesting to learn how IoT sensors can transform railway switch maintenance and what it all impacts with one little piece of technology for such a critical piece of infrastructure. Thank you so much for sharing all your insights and your knowledge. I hope to have you again on the show, maybe in a couple of years where we then can look back and wonder why we put up with the bad infrastructure for so long.

Oliver: Looking forward to it. Yes.

Ruth: Thank you so much for listening to We Talk IoT. Until then, keep innovating and stay connected.