Drone Trains Won't Stop Wrecks

If you were trying to crash a train, as in, intentionally throw it off the tracks and kill people, pretty much your only option would have been to do what Metro-North engineer William Rockefeller did on Sunday.  What happened, or appears to have happened, is the rough equivalent a pilot pushing an airliner into a terminal dive or forcing a catastrophic stall. But blaming mechanical failure is much, much easier in aircraft crashes. The failure needed to send this train into a tight curve fast enough to toss it off the tracks against the actions of the operator would have to be extreme. Instead, it looks like the engineer zoned out hard in one of the worst possible places on Metro-North's Hudson line.

It turns out that New York Governor Cuomo was most likely on point when he threw Rockefeller under the bus barely 12 hours after the accident. “It was a tricky turn on the system, but it is a turn that has been there for decades,” he told the Today show. The train hit the curve, rated at 30 mph for safe operation, doing 82 mph.

Rockefeller cut the throttle six seconds before the train derailed. That works out to just over 700 feet before the train spilled. The curve, poised at the end of a straightaway stretching north practically to Yonkers, hogs most of those 700 feet, meaning the engineer took his shit likely less than a second before the train hit the damning curve.

MTA spokesperson Marjorie Anders confirmed to me today that there would be no warning in the train’s cab as it approached the curve going too fast. The train’s defense against what happened Sunday was strictly Rockefeller and his response to both the looming curve itself and, right at beginning of the curve and not before it, a speed limit change. Had he not derailed, Rockefeller could have blasted onward through any speed-zone without any warning beyond a speedometer in front of him. A pilot has stall warnings and altitude warnings to ignore. On a Metro-North commuter train, the alarms only kick in when a train goes past a red signal or is approaching another train. It does nothing if the train is barreling toward a cliff disguised as a rail curve

There is indeed a bit more nuance to throwing a train off of a curve than crashing into something on the tracks in front you. A derailment is not merely a crash, but a more subtle work of physics that’s worth understanding. The next few weeks/months will see a lot of calls for something “positive train control,” which is basically a way to control a train remotely and overriding a possibly negligent or dead operator. The train becomes a GPS blip of data plugged into software that might slam on the brakes in any number of different situations, including possibly an engineer not reacting properly to an impending curve.

The Long Island Railroad, just right over there and operated by the same agency as Metro-North, has a proto-positive train control system in place that slows trains down if they break the speed limit or approach another train. And a law passed just after 2008’s hellish Metrolink disaster mandated all railroads in the United States implement PTC systems, but the rollout has been spectacularly slow. And a 2012 Federal Railroad Administration report found that most railroads won’t be able to comply with the law by its 2015 deadline and recommended extending that deadline indefinitely. It also found that PTC technology is only likely to prevent up to 30 percent of rail accidents in the first place.

Turning trains into drones has other problems, which should be fairly predictable in the age of Stuxnet and growing fly-by-wire infrastructure. A system that can slow a train down before a dangerous curve can, in different hands, make it go faster into that curve. In some quarters it's thought that a hacked command-and-control system moved a Blue Line 'L' train with no operator present.

In the end, no technology could have saved Sunday’s doomed Metro-North train. However late, it’s almost certain that Rockefeller acted faster than a positive control system because such a system would still only be bound by the speed limits in place on the line. When the 30 mph speed zone might have triggered his train’s control systems at the very beginning of the curve, he was already trying to stop. And it was already much too late for that.

A recent GAO report is much more enthusiastic about at least one train safety idea: building better railcars. That is, cars that crash better and derail safer. Experts consulted in the report were enthusiastic about building both better and safer fixtures in rail passenger cars (seatbelts, perhaps?), and also some rather high-tech ways of managing the energy involved in a crash. Amtrak's brand new ACS-64 engines come with new coupling mechanisms designed to prevent rail cars from flipping or jack-knifing in a derailment or collision, two things with huge implications for an accident's survivability The new. locomotives are also designed to absorb a 500,000 pound hit (about two stopped locomotives), taking at least a significant amount of shock away from the more vulnerable passenger cars.

The Metro-North train involved in Sunday's accident was, relatively speaking, old and busted. Pushed from behind by a diesel engine, the driver sat in a small cab at the exposed end of a passenger car. Reversed, the train would at least have the front-end buffer of a locomotive. Though, in fairness, any separate engine or engines in a modern commuter train are becoming rare in favor of cars that double as engines either as diesel-multiple units or electric-multiple units, like those seen on many other LIRR and Metro-North Lines. The passenger cars involved in Sunday's crash were most likely Shoreliners, a popular and cheap model that hasn't seen much of any updating since its 1983 introduction. Maybe it's time.

@everydayelk