Bumped Down: The 2007 Tarvisio-Boscoverde (Italy) Runaway Train Derailment

Max S
10 min readSep 26, 2021

--

Background

Tarvisio (also called Tarvis) is a city of 4140 people (as of December 2019) in the extreme northeast of Italy, located 57km/35.5mi north-northeast of Udine and 9km/5.5mi southwest of Arnoldstein (Austria), which is as far north of the Austrian/Italian Border as Tarvisio is south (all measurements in linear distance).

The location of Tarvisio in Europe.

Tarvisio-Boscoverde station is a large train station to the east of Tarvisio itself. It is the “split” between the Austrian and Italian railway network, meaning this is where electric locomotives and multiple units have to switch from one electrical system to another. Austria uses 15kV AC while Italy uses 3kV DC on nearly all their lines. Modern trains passing through the station can do this switch quite easily, they will lower the pantograph at a marked point, “coast” through an isolated section with no power on the overhead wires, set their systems for the new electrical system and raise the pantograph up to the wire again. Alternatively a locomotive will be parked with the brakes released and be pushed or pulled to the other section by a different locomotive. The same thing is done if a locomotive, especially one that can’t use both systems, is to return to where it came from. Stopping at the station usually means rolling into the opposite system, so a locomotive on the new side will push the locomotive in the opposite direction to give it enough momentum to roll back to where it can work with the overhead wires.

The northern end of the station is the southern end of the Rudolfsbahn (“Rudolf Railway”), an electrified mostly double-tracked main line connecting the Italian border with St Valentin in northern Austria via 407km/253mi of track. The line opened in sections between 1868 and 1873 and has been modified several times since then. Its full expansion is mostly used for freight trains, with a few passenger services using the line within Austria. Nowadays the trains on the line can go as fast as 160kph/99mph. It’s one of 4 north-south routes across the Alps, alongside the Semmering railway, the Tauern railway and the major Brenner railway.

Track 5 at Tarvisio Bsocoverde station seen from above, the locomotive rolled away towards Austria (upper edge of the photo).

The train involved

ÖBB (Austrian national railway) series 1116 number 062 was a Siemens ES64U2 “Taurus”, a four-axle electric multipurpose locomotive designed for international trains. The 1116 specifically is a version of the 1016 with added compatibility for the 25kV-system in Hungary. Each 1116 measures 19.28m/63ft at a weight of 86 metric tons (since the accident increased to 88 metric tons by added equipment). They have a power output of as much as 6400kW/8700hp (under 15kV) and can reach as much as 230kph/143mph. At the time of the accident 1116.062 had been rented out to the GySEV (Raab-Oedenburg-Ebenfurter Eisenbahn), a Hungarian-Austrian railway company largely owned by the Hungarian government and wore large decals covering up the ÖBB’s logos. The GySEV owns 5 Taurus-locomotives (numbered 470 in their system), recognizable by their yellow-green livery, and had rented another six (including 062) from the ÖBB at the time of the accident. At the time of its introduction in 2001 a series 1116 cost approximately 2.63 Million Euros/3.08 Million USD.

ÖBB 1116.062, the locomotive involved in the accident, photographed with GySEV-decals in 2006.

The accident

On the 18th of May 2007 at approximately 10:10am 1116.062 arrived at Tarvisio-Boscoverde station on track six. It had helped ÖBB 1216.002 (a Mark 2 Taurus-locomotive) pull a 1005 metric ton freight train up the alps and across the border. 1116.062 was meant to be disconnected, turn around and head back to Villach in Austria on its own while 1216.002 would be taken over by an Italian driver and pull the train further into Italy to Cervignano. As the train came to a stop at the station the driver was met by an Italian shunting worker whom he handed the train’s paperwork. Wrongly assuming that both locomotives were to return to Austria together the driver left the locomotive with the brakes off (to be moved back to the Austrian electrical system’s side of the station) and, as it would be the rear locomotive if the two headed back together, set the locomotive’s system to “slave” mode. This disabled the locomotive’s signaling system component, meaning the signaling system would treat it like any train car, not like a locomotive. Afterwards the driver went to the rear cab of 1216.002, assuming this would be the leading cab for the drive back, and prepared it to be pushed over also. He had simply forgotten that he was supposed to hand off 1216.002 to the new driver and return with only 1116.062.

1216.002, the rear locomotive of the freight train, photographed in 2017.

Meanwhile the Italian shunting crew, who had the correct information, disconnected the two locomotives and handed the data-cable that had been connecting the two locomotives to the Italian driver who had entered the forward cab of 1216.002, unaware that his Austrian coworker was at the other end of the locomotive. When the Austrian driver noticed a shunting locomotive pulling 1116.062 being pulled away from the train he radioed the ÖBB’s office at the station, telling the liaison worker that the locomotive was without brakes, in slave mode and unoccupied. The liaison worker forwarded this information to her Italian colleague in charge of shunting operations, in both German and Italian, who told her that it’s okay and that they will wait. However, for unknown reasons this decision was not forwarded or obeyed by the shunting crew, who finished pulling the locomotive out of track six and, after disconnecting from the shunting locomotive, pushed it into track five to make it roll back into the section electrified with the Austrian system. Protocol dictates that before a locomotive is “pushed off” (separates from the shunting locomotive) radio communications with the driver aboard the pushed locomotive have to be established and the driver of the pushed locomotive is actually the one who gives final permission for the maneuver. None of this was done, obviously, as there wouldn’t have been anyone aboard 1116.062 to contact in the first place.

When the Austrian driver aboard 1216.002 saw “his” locomotive roll towards Austria on the neighboring track he again radioed the liaison worker, again telling her that the train, which is now moving on a different track on its own, is unoccupied. Seeing that a conversation via radio had failed the worker left her office and went to her Italian coworker’s office to forward the information in person. It was 10:38am when she reached the other office, just as the station’s traffic control system reported a set of points on the Austrian side being forced to shift by a rail vehicle going through them in the wrong direction. 1116.062 had gotten out of control, and was now headed for Austria on a downhill track with no brakes or driver. The crew at the station knew they couldn’t go fast enough to catch it with another locomotive, much less safely couple up to it at speed. Shifting to emergency protocols the liaison worker checked the path the runaway locomotive was travelling on and informed the dispatcher at Arnoldstein station, approximately 9km/5.5mi down the tracks from Tarvisio Boscoverde station. He in turn informed the signal box worker at the ÖBB Terminal Villach Süd, a large freight yard 7km/4mi beyond Arnoldstein station. An express train towards Italy had departed Villach just minutes ago, and while the radio call came too late to stop it at Villach Süd Villach’s signal box worker managed to contact the driver of the express and tell him to hold on a siding at Neuhaus station halfway between Villach and Arnoldstein. Had the express passed Neuhaus it would’ve likely collided with the runaway locomotive speeding towards Villach. One does not want to imagine what the outcome of that would’ve been like.

A railfan takes this photo of the unoccupied 1116.062 speeding down the tracks just after crossing into Austria, note how both pantographs are down.

Knowing that the signaling system is useless in stopping the locomotive the decision is made to divert it into the freight yard at Villach Süd. If the locomotive somehow navigates the sharp S-turn at the entrance to the freight yard workers can try to derail it inside the freight yard, create a controlled collision with freight cars or perhaps slow it down with brakes intended to hold locomotive-less freight cars. If all else fails it can be diverted into a sharp 180° turn at the other exit of the yard. Either way the locomotive won’t go into the towns of Villach or Neumüllern, where it would endanger residents if it derailed.

The set up path to get 1116.062 and where the express train was stopped.

Due to the steep downhill grade of much of the track north of Tarvisio Boscoverde station the locomotive reaches as much as 158kph as it passes through Neuhaus station (where the express train was placed in a siding to let the runaway locomotive pass). With the track leveling out and the upcoming S-turn the locomotive looses some speed as it turns off towards Villach Süd. Any preparations to slow or stop the train at Villach, preferably intact, are for nothing though. The locomotive makes it through the left hand turn, but the following, tighter right hand turn is too much. At 10:41am the locomotive derails right outside the freight yard, falling out of the turn at 135kph. The locomotive rolls over as it falls down a small hillside towards the Gail River, crushing a support pole for the overhead wires and tearing down several trees before coming to a rest on its side just about 150m/490ft from the tracks. The locomotive is severely damaged, a large cut has been torn into the woods and 2000l/528gal of transformer oil need to be cleaned up, but no one got hurt.

The locomotive after being recovered from the site, showing the severe damage caused by rolling over (note that the cab roof was removed after recovery).

Aftermath

Operations at Tarvisio Boscoverde station largely proceed as usual, aside from investigators interviewing the involved employees and reviewing the day’s paperwork. The accident shuts down the adjacent freight yard, the overhead wires are destroyed and approximately 100m/330ft of track need replacing. The fire department cuts down some more trees and removes some of those the locomotive had mowed down, the day after the accident a crane is brought in to upright the heavy lump of steel. A few days later two rail-cranes are used to remove the wreck from the site and take it to a maintenance yard.

1116.062 being uprighted the day after the accident.

It becomes clear fairly fast that the locomotive was in perfect working order, as was the signaling-system, and that human error is to blame for the accident. However, the investigation also points out that, once things were (literally) set in motion everyone involved worked perfectly as trained, enabling clear communications and constant surveillance of the runaway locomotive. This also allowed the passenger express train, which was coming the other way on the same track (as 1116.062 had become a “wrong way driver” of sorts) to be directed to a safe waiting position. The accident sparks discussions about the safety of the “push-off” technique, which isn’t as widely done anymore as it used to be. It still is widely done at some stations, including Tarvisio Boscoverde or the larger Brenner pass station. Some discussions literally bring up the argument of “because it’s fun”, while objective reasoning makes it more likely that the actual reason is that everyone knows how to do it and that there is no real incentive to change what worked fine for decades. In the end the accident is blamed on a fatal breakdown of communications, causing two different schedules to circulate (one verbal one actually printed out), along with negligent, rushed behavior by the Italian shunting crew who proceeded with the operation despite (obviously) no driver communicating with them from aboard the locomotive. The report lists a number of recommended improvements, essentially asking the ÖBB to scrap the current guidelines for push-over operations of locomotives. They recommend that locomotives are not to be set to slave-mode for the operation and just “bumped and let go” but instead must be completely hooked up to the shunting locomotive and moved in a controlled fashion up to a complete stop. Employees are also to be reminded to never even start the maneuver without proper communications being established. Lastly, the ÖBB and FS (Italian national railway) are advised to improve the foreign language proficiency of their employees especially regarding words an terms involved in railway operations to reduce the risk of misunderstandings.

The remains of 1116.062 photographed in August 2009.

After the accident the ÖBB officially announces that the locomotive will be returned to service, in the meantime GySEV is given 1116 060 to still fulfill the rental-contract. In the end some internal parts are stripped from the severely damaged locomotive before it’s written off and set aside for scrap. It is in that queue (of sorts) where a railfan spots the sorry remains of the locomotive in 2009 (see above), sitting next to the remains of the Taurus destroyed in the 2007 Szőny Train Collision. The ÖBB had ordered 3 spare body shells from Siemens earlier in 2007, one of which is used to rebuild 1116 062 (now technically 1116 062 II), however, how much of the original remains in the new one is unclear. The new 1116 062 II was first spotted back in service for the ÖBB (as the GySEV had number 060) in January 2011.

The rebuilt/new ÖBB 1116 062 II pulling a “Railjet” express train in summer 2011.

_______________________________________________________________

Join the discussion about this post on Reddit!

The index of all my posts on Reddit.

--

--

Max S

Train crash reports and analysis, published weekly.