What does it take to make the trains run on time?
It may seem like it's time for a history lesson here. In Silicon Valley, after all, where everyone is thinking about new self-driving cars, sometimes we forget how it all began: the railroads. But even today, a large chunk of the American economic engine is driven by classic railroad technology. The industry drives around 1 million additional jobs, accounting for a third of American exports, according to data from the American Association of Railroads and the U.S. Department of Commerce.
That's why it's a badge of pride for a piece of industrial engineering technology to find a home in the world of trains. It's even more important that this tech which offers the country so much continues to be safe and reliable in the 21st century.
There's one major issue at hand when it comes to engineering "smart" trains: how do we make train cars communicate with one another? That inter-communication between train cars and a train conductor is crucial while the train is moving, as it sways along tracks which are sometimes older than we realize; but it's also especially important for braking. Trains are, of course, often miles long. A human conductor in the front or even middle of the train can't be expected to know what's happening in the event of a fault or sputter two or three miles behind at the caboose.
So whatever technology makes the cars communicate has to account for both their moving and stopped states - not to mention the unique mix of locomotives and cars that comprise any train.
For Dale Stevens, a manager of software engineering at New York Airbrake, a company building innovative and integrated tech solutions for modern railroads, there’s one central, behemoth of a challenge to think about when it comes to train tech:
"It's dynamic. It's moving," he says. "You'd think of a more typical installation as more of a static environment - a building or power utilities system, for example. You kind of go install the equipment, commission everything and you're good to go until you have to replace everything. It's not that simple with trains."
So, thinking back to that two-mile-long train, what's needed is a system that can rapidly and reliably send a slew of commands from the conductor down through all the cars, doing so by way of cables, connectors, and junction boxes. That's where New York Airbrake chose LonWorks technology to solve a braking-tech problem that's existed in train cars since their inception in the early 1800s. Not until the 1990s did train engineers begin imagining a world where electronic braking was possible - something we can now take for granted on our freight systems.
"Sometimes we forget how it all began: the railroads. But even today, a large chunk of the American economic engine is driven by classic railroad technology."
What's key for Stevens about LonWorks technology is the fact that it can depend on existing power-lines. Imagine a world where trains had to essentially carry their own telephone poles with them on a track from New York down to North Carolina. Weight and bulk would reach absurd and untenable levels. Instead, the train communications can use the power lines surrounding the train - an amazing feat considering, again, that the train is moving this whole time, and therefore always needing to adapt to its surroundings.
Then Stevens also has to think about the line-up of the cars; he can rattle off about fifteen possible train configurations in one breath (three locomotives in front; hundreds in the back, or maybe two up front, more out back, or maybe two then a few cars and then three more locomotives, etc). That should offer some perspective on just how many - truly infinite! - possible layout combinations the New York Airbrake team has to think about when designing their braking solutions. Yet with all of that, they're still able to send stable and strong messages in good amounts of time. In a train with about 100 cars, Stevens says, it will take about 100 seconds for a message to whiz down the line, through all relevant parts.
Stevens says in the train world, that speed is enough. Things are kept at low bandwidth - this isn't a system that needs rapid fire communication, but rather depends on a few key messages broadcasted out. How it works: the lead locomotive sends a broadcast, which is captured by the length of train following it, using those transceivers attached to each train car. Then the lead locomotive can capture the status of each of those cars via a signal (or in case of emergency, an alarm) traveling back in the other direction. Quickly, the train engineer may be able to notice that, hey, there's a problem a mile back. But rather than burdening the software with excessive message technology - an unnecessary, bandwidth-consuming possibility - these train cars are built to "hear" rather than "talk to" one another, strictly speaking, Stevens says.
"Putting electronics on freight cars is new to everyone," he said of the challenge of bringing technology to a bulky industry. "You go from being able to operate maintenance out of a totally wrench-based shop to needing meters and protocol analyzers."
But that change is eased with technologies like this one, which don’t attempt to suddenly burden last decade’s technology with solutions built for next decade-grade stuff. From the outside, progress seems to happen in a flash, but within the industry, progress is all about how good the next-engineered feat can be. An enormous part of that slow crawl of progress is the challenge of building dynamic-interfacing, backwards-compatible, comfortable technologies for industries like freight, on which so much of America depends.
With this technology in hand, however, it’s an exciting time to think about railroads and all the little engines that could.