Production of next-generation Acela Express fleet underway
Stadler unveils TEX Rail Flirt DMU
Siemens invests in remote monitoring specialist Wi-Tronix
DB consortium selected for California high speed rail
Judge puts the skids on state’s proposed rail trail
Amtrak's CEO shares his vision for rail's future
Flight Rail: a new type of train?
America’s short lines play the long game
New York rail operator bolsters security after London bombing
FreightWaves recently chatted with Tom Forbes, vice president at Navis Rail, a technological solutions provider for the freight rail industry.
Forbes was formerly CEO of Melbourne, Australia-based SaaS provider Biarri Rail, which Navis acquired in February. Navis itself is part of Cargotec, a Finland-based provider of cargo and load handling technological solutions. Navis provides technological tools for improving supply chain flows.
FreightWaves and Forbes talked about what role artificial intelligence and big data should have in freight rail and within the overall supply chain. This Q&A will be in two parts.
This interview was edited for clarity and length.
FreightWaves: What do AI and big data look like in the freight rail industry?
Forbes: Let’s start with big data because big data [large volumes of data] is a phrase. … What’s interesting to me in freight rail is that data collection has been going on for years in terms of the instrumentation. The wayside and signals and all of those types of things have been around for a long time. And we’ve seen the [use of] sensors increase exponentially over the past five or so years, particularly on the rolling stock itself, locomotives, wagons, as well as by the wayside.
So rail has a lot of data to analyze, but that data has been analyzed in a kind of real time or static sense. A sensor goes off and tells me now about something that I need to react to now.
What’s been happening over the last five years is that we’ve started to say, “Let’s collect that data over the last five years and see what happens over time rather than in the moment.” And so, the first wave of big data/AI, machine learning has been the use of predictive technologies. If you can tell me about how something performs historically — and particularly how it fails — then you can start to predict failure based on a whole range of attributes and variables.
That was the first wave of big data, which was to say, “Huh, this is cool! We can use this data collected over time for predictive analytics.” And that continues to happen more and more.
As we have more connected sensors, we’re collecting this data and people can do predictive analytics and see big wins around those things like safety, reliability and efficiency, condition-based maintenance. Keep in mind that rail is a capital-intensive business, so anything you can do to improve the efficiency or reliability of those assets is going to have a big impact on your railroad.
Looking into the future, I think what’s exciting — the surface has barely been scratched on this, quite frankly — is you can improve the operational efficiency of what you’re doing: deciding how to operate, what to operate, when to operate, the planning and scheduling function of a railroad. … [Through using] all that data and understanding how things perform, we can build an operational plan that is more robust and more in tune with the historic performance of what we know about things like seasonality.
It’s not building operational plans in this laboratory condition where you say, “All things being OK, we’ll build a plan and we’ll operate and we’ll be fine.” We know that will never happen. We know that things will never be fine. The problem is we’ve never had a good way of knowing.
But now we can imagine and predict and use this data to have a better idea of how it will be fine and how things will vary. We can build our plans with flexibility and robustness to deal with a known level of variability rather than an unknown level of variability. … We can make better assumptions and use big data to inform the algorithms of artificial intelligence to come up with plans based on a level of variability that’s more predictable.
And I think that’s really cool. It’s a step change in the way people think about planning and operating their railroads.
FreightWaves: Should the freight rail industry be setting short-term and long-term goals on how to integrate more of the AI piece overall?
Forbes: It’s an interesting question. The short answer is yes, probably. The longer answer is there are a couple of things happening simultaneously, and they’re not mutually exclusive. There’s an interplay between … automation and connectivity.
Automation in rail is a bit different than automation in trucking. Automation in rail isn’t all about how do we handle an autonomous train. That’s a small portion of it. Think about a larger train and how much more freight it’s moving [than a truck]. Trucks are in order of magnitude smaller.
Automation for removing a trucker is a bigger net gain for the trucking industry than it is for the freight rail industry. … Automation in trucking is a threat to rail in terms of overall efficiency and competitiveness, but the response of the rail industry [shouldn’t be] to do automation to get the same result — i.e., remove the engineer or the driver. It’s for other efficiency reasons.
It’s interesting because I had a conversation with some folks at Rio Tinto in western Australia. They’ve got iron ore trains that have been the pinup for railway automation. What almost surprised them was that the real benefit was operational flexibility — not to have to worry about train drivers. Train drivers have restrictions of hours, of breaks, changing drivers that slow the train down — all the things you have to do that change the way you operate a train because the driver is present. Yeah, removing the driver is removing the cost, but the ability to run trains with more fluidity, to run them through the network in a more automated way — for North American Class Is, for example, velocity doesn’t get talked about as much as it used to, but that velocity is the key for any rail operator. … That’s the kind of automation holy grail for rail, to have this fluidity to your network.
Containers and railcars in a yard. (Photos: Jim Allen/FreightWaves)
The other trend that I think is important is connectivity between parts of the supply chain. This is less of a trend that people talk about. It doesn’t get the headlines. But I think it’s equally important. This has been a little bit of an eye-opener for me as Navis acquires [other business products] and we do more work between what’s happening in yards and terminals and rail operations.
To some extent, the integration of Navis and Biarri Rail has been a microcosm of the industry. … We’ve had to educate each other. Folks with Biarri Rail have really had to talk about how rail operates in rail yards, and folks at Navis have had to educate folks at Biarri Rail about how terminals operate and the visibility they have for vessels and containers and how they’re getting loaded onto rail. And I really think it’s interesting that both of us had almost a blind spot between those two points.
But that’s true of the industry in general, the kind of visibility from the beneficial cargo owner or the shipper or the ocean carrier that’s moving the freight to the terminal operator, which might be a rail operator, to the rail and, of course, trucking is in there as well. That’s broken. Those interfaces don’t work very well currently. The terminal operator says, “I wish I knew where the train was.” And the train operator says, “I wish I knew when the containers are going to be there to load.” … And that’s not for lack of data … [but] the data is not connected. That connectivity of data has a lot of value of efficiency for every stakeholder in that chain.
If you ask the question, “Why? What’s the problem?” Part of that comes down to just that it’s hard to coordinate multiple stakeholders. … But some of it is the data interchange and the extraction of that data, things like blockchain. I must admit that I didn’t understand quite a few years ago where blockchain fit in the supply chain. But this is a really obvious use case, right? How do we have data that can be transmitted securely between all the various stakeholders so that it can be used to provide efficiency gains? There’s a real opportunity there. And it doesn’t have to be blockchain, but I can think blockchain can be a real enabler in the safe and secure transmission of data between all the stakeholders. Add automation, and in my mind, those two things together are very, very powerful.
Tom Forbes of Navis Rail. (Photo: Navis Rail)
FreightWaves: What lessons have you learned from terminal operations that could be transferred to rail?
Forbes: One of the interesting things for me at least in my learning about how those areas work is this issuance of instruction to equipment, and as I mentioned earlier, terminals are automated much more than rail operations.
Even before you go into the railway network, I think there’s an opportunity for rail operations to modernize their rail yards more significantly along the lines of intermodal terminals … [where] operating systems issue instructions that are automated or semiautonomous or even to manually move equipment. … I think rail operators with rail yards could push into that fairly rapidly, following what’s already been done.
FreightWaves: What should the federal government’s role be in encouraging or regulating AI in freight rail or transportation in general?
Forbes: I think a lot of people think regulation stifles innovation, and I can certainly understand that viewpoint. But I think governments can have a role that is on a spectrum, from strict regulation to things that are less restrictive and more beneficial, such as setting standards.
I mentioned earlier blockchain and direct chain, and I’ve talked about exchanging data beyond the supply chain. That’s a really obvious example of where government can start to provide, by encouraging collaboration and working with stakeholders around things like standardization, classification, quality standards, sharing standards, protocols, those types of things.
And to some extent, technology providers will find themselves often in the sandwich between stakeholders having to perform that role. … But I don’t think the technology vendor should be setting standards more broadly. The risk there is that there will be solutions that are fragmented because each vendor or each stakeholder tries to solve it themselves.
If I’m running a railroad or terminal or whatever and I’ve got problems today, I don’t want to wait and have government sort it out. … But those problems in the longer run need solutions because the nature of the supply chain is global. If you’re a vessel operator, you don’t want the terminal standards in the U.S. [to not align with the ones] in China. It might be fine for a rail operator, but again, if you’re a rail operator and you’re running trains through Canada, the U.S., Mexico or, even worse, Europe, where you’re running across a lot of countries, these kinds of standards are going to become really important. That’s the role that I would like to see government playing — to encourage and collaborate with industry to help set those standards so that we don’t end up with a really fragmented world.
FreightWaves: That reminds me of how positive train control has been implemented here.
Forbes: That’s spot on. That was obviously originally driven by a safety need but the federal government realized that PTC needs to be across freight, passenger — across all rail. And it needs to be interoperable for it to work. That is actually a really good analogy … but the difference is that it’s not just an American problem, it’s a global problem. And so that’s going to require an even broader collaboration across a set of governments and industry.
FreightWaves: How has the COVID-19 pandemic affected freight rail?
Forbes: It’s hard to have a conversation like this and not talk about COVID. But one thing I would say is, for me, COVID has been an exclamation point on an increasing trend of having volatility in the freight markets.
Volatility has always been before us. We saw it through the Great Recession … and then coal going hot and then off, fracking and frac sand, particularly in terms of freight rail, and oil of course — all of these things that have changed and evolved and kind of caught industry on the hop a bit. And of course, COVID has been the latest example of that, where volumes did come back but they came back in different ways.
The rail industry has been around for so long and it’s operating in a pretty methodical way, but it all just highlights this increasing need for organizations — not just railroads but anyone involved in the supply chain — to be much more nimble and reactive to events and to be able change how they operate kind of overnight. I was actually quite impressed with the way the rail industry did react, but we saw awaken with our clients this increased need to say, “Oh, I really need to rely on technology more than ever, more than the historical reliance on the industry expert whose been in my organization for 30 years and whose seen and all and knows it all.” Even they don’t know what to do when something like COVID happens.
And even those industry experts — they’re retiring and you’re not getting that same railroader-for-life employee anymore — the work dynamic is changing. And so with that expertise, that institutional knowledge reducing at railroads over time, along with this increasing volatility, the need to rely on technology — and particularly data, decision-making AI type of stuff — is just becoming ever more important. COVID’s not a wake-up call because I think this issue has been recognized, but it’s certainly really highlighted that need across the supply chain.
This article first appeared on www.freightwaves.com
About this website
Railpage version 3.10.0.0037
All logos and trademarks in this site are property of their respective owner. The comments are property of their posters, all the rest is © 2003-2021 Interactive Omnimedia Pty Ltd.
You can syndicate our news using one of the RSS feeds.