Accidental Automation: The 2013 Vienna-Penzing (Austria) Train Collision

Max S
9 min readMay 12, 2024

Background

Penzing is a former municipality and now district of Vienna, Austria, counting 2187 citizens as of 2020. The district is located on the far western outskirts of the Austrian capital, 8km/5mi west of downtown Vienna and 49km/30mi east of St. Pölten (both measurements in linear distance).

The location of Penzing in Europe.

Penzing has a station on the “Vorortelinie” (“Suburban Line”), an electrified, partially dual-tracked commuter rail line on the western side of the city. The line opened in 1898, originally mostly serving freight traffic, before eventually focusing more and more on passenger services. It eventually became part of Vienna’s S-Bahn, an urban and sub-urban commuter rail system, in 1987, being designated as S45. The line can be used at up to 100kph/62mph.

The site of the accident seen from above, with the adjacent stations Hüttendorf (left/west) and Penzing (right/east) visible towards the edges of the image.

The Trains Involved

Both the westbound service from Wien Handelskai station to Wien Hütteldorf station (Z 20592) and the matching eastbound service (Z 20595) were provided by an ÖBB (Austrian national railway) Series 4024, with 4024 101 travelling westbound and 4024 129 travelling eastbound. The Series 4024 is a four-car electric multiple unit (EMU) manufactured by Bombardier Transportation as part of their “Talent”-family of trains. The type was introduced in 2004, along with a three-car version (4023) and a four-car model for services to Hungary (4124). Each 4024 measures 66.87m/219ft in length at an empty weight of 139 metric tons* and can reach 140kph/87mph. The four-car version offers 199 seats in a single-class configuration along with space for another 252 standing passengers. It’s not known how many passengers were on either train at the time of the accident, although some media-outlets claimed that the trains were “full” or even “crowded”.

*This is the weight listed by the official report on the accident, other sources give weights as low as 116 metric tons

An ÖBB 4024 identical with those involved in the accident, photographed in 2012.

The Accident

ÖBB 4024 129 reaches Wien-Hütteldorf station from the east at approximately 8:30am on the 21st of January 2013. The driver switches to the other cab after all passengers have disembarked, preparing to drive the eastbound service Z 20595 to Wien-Handelskai. The signal box pre-sets the train’s eastbound path in accordance, but the train can’t depart as the signaling system refuses to turn the exit-signal (H 106) green, claiming that the adjacent section of the rail line is occupied. The dispatcher in charge of Wien-Hütteldorf station consults the logbook and his colleague in charge of the neighboring Wien-Penzing station, arriving at the conclusion that the system is displaying a faulty occupancy status. He thus uses a replacement signal to override H 106, allowing Z 20595 to depart past the red signal. The dispatcher, having solved that issue for the moment, turns to other tasks as Z 20595 quickly accelerates out of the station.

The train passes through Hütteldorf’s freight yard and turns a left hand corner when, just over 600m/2000ft into the train’s journey, the driver spots the headlights of another train in the distance, on the same track he is on. Both train drivers trigger an emergency stop, but their trains are already too close to come to a stop. They thus collide head-on at 8:43am, at a combined speed of 75kph/47mph.

The westbound train, 4024 101, digs itself into the oncoming train’s nose-section at 49kph/30.5mph on impact, managing to bend up the other train’s forward floor as it obliterates the driver’s cab. Z 20595’s driver had abandoned his seat moments before impact, allowing his survival. He is among 9 severely injured survivors, his coworker on the other train and another 40 passengers suffer minor injuries in the collision.

Aftermath

Various people alerted the emergency services moments after the collision occurred, allowing a large number of responders to quickly reach the site of the accident. The location of the wreckage atop a high, snow and plant-covered embankment still slightly slows down the access to the train. Firefighters help the injured survivors down the embankment to ambulances waiting in the street below, while ÖBB-employees guide uninjured passengers along the tracks to nearby stairs that lead down to street-level as well. The police, the ÖBB’s own incident investigators and those from the SUB (Federal Safety Investigation Department, the Austrian entity in charge of investigating accidents involving trains, cable cars, civilian aviation and ships) take over the site once the trains are cleared, trying to figure out how two trains ended up colliding on a single-track line despite the safety-measures meant to avoid exactly this.

The front end of ÖBB 4024 129 after the trains were dragged apart, lacking essentially its entire cab-section.

Suspicion soon falls on the dispatcher for Hütteldorf station, who had used the replacement signal to override the red signal preventing Z 20595 from departing on its eastbound journey. The use of replacement signals has a tragic history regarding its usages ending in catastrophes, as they are a manual way to override a vital part of the signaling system. Modern signaling systems automatically detect the positions of each train running on their tracks and will keep signals red for other trains if the track-section beyond the signal is occupied by a train. But since any system can be defective dispatchers still have to be given a way to override the system, an ability that was coupled to a specific process to ensure maximum safety if the signaling system had to be overridden. The dispatcher who had activated the replacement signal was new on the job and still working under supervision, further feeding suspicion that he may have activated the signal without following proper procedures.

The SUB investigators talk to both the acting dispatchers and the supervisor watching over the actions of Hütteldorf’s trainee dispatcher, but conclude that all procedures and guidelines had been properly obeyed. The supervisor explains that an erroneous emergency stop by an earlier train had caused the signaling system to develop a faulty occupancy-status for the single-track section between Hütteldorf and Penzing. That train’s driver had been routed via the neighboring track, which belonged to a different rail line, due to an accumulated delay which would have otherwise further delayed Z 20595. The driver triggered an emergency stop when he realized this, reporting that he wasn’t certified for that rail line (which runs directly parallel to the Suburban line the whole way).

The trainee thus correctly switched to the abnormal operation mode for this exact case, which involved individually “offering” trains to the neighboring stations and confirming each train’s arrival, too. The trainee, under observation by the supervisor, perfectly followed the whole procedure, ending with a correct activation of the replacement signal on H 106.

An Austrian main signal showing the replacement signal, a flashing white light, allowing the red light to be disregarded.

The driver of Z 20595 was also interviewed once he had recovered enough to do so, and his statement aligned with those by the dispatchers. He had spotted the flashing white light on the red main signal as he approached it, acknowledged it as required by pressing a button on his control desk and then proceeded past it at the standard speed limit for such a situation of 40kph/25mph, only accelerating further once he had left the immediate area behind (allowing him to do so). He had almost reached the local speed limit of 80kph/50mph when he spotted the oncoming train approximately 400m/1300ft in the distance, leading to an emergency stop being triggered before he retreated into the passenger area right before impact.

The leading ends of both trains after the accident, with ÖBB 4024 101 sitting noticeably higher as it had dug itself into the other train.

The core cause of the accident thus didn’t lie with the replacement signal but with Penzing station, where operations were perfectly normal. Which, as it turned out, was exactly the problem. Dispatches at Penzing station were done via the so-called autonomous setup operation (“Selbststellbetrieb”), which essentially means the computer system set paths, points and signals by itself with the dispatcher mostly observing during regular operation and stepping in when needed. The system was meant to be turned off when operations were switched to abnormal operation procedures. The dispatcher hadn’t forgotten to do this, but happened to do so exactly when his coworker for Hütteldorf provided input into his control desk which was automatically fed to the system, which prioritized automatic inputs over manual ones and thus disregarded the “turn off”-order. This part, the fact that the system hadn’t shut down as instructed, went unnoticed by the dispatcher.

Part of the departure-preparation for the eastbound Z 20595 involved pre-setting its path ahead of its signal being turned green, which occurred before it became clear that the signaling system had developed a fault. That process involved switching the operational direction of the line from westbound to eastbound. This switch would illuminate an arrow on the dispatcher’s control desk, pointing to the right (east), but the system would actually only adopt the status once a train was dispatched.

An (enlarged) photo from the report showing the indicator at the control desk for Hütteldorf station as it would have presented itself to the dispatchers.

Normally the dispatcher would have then received confirmation from the system that the path was clear and the signal would then be turned green as the dispatching train concludes the system switching directions. Things never got that far on the day of the accident, though, as the system rejected the pre-set path due to the faulty occupancy status. H 106 was never turned green, instead being overridden with the replacement signal, which meant the system never adopted the switch from westbound to eastbound operation as there had been no regular dispatch. The autonomous setup operation at Penzing station thus continued to register the section as “westbound” and thus allowed the dispatch of Z 20592. The dispatcher at Penzing realized the mistake as the train departed the station and sent an emergency stop order, but the driver had already triggered the emergency stop by that point, which was still too late to avoid the collision 1.2km/0.75mi beyond Penzing station.

The inside of 4024 129’s forward section after the accident. Some of the debris, like the red and yellow buttons, belonged to the driver’s desk (the large blue piece).

The report finds the main cause in the autonomous setup operation being left on during abnormal operations, but notes that the systems available to Penzing’s dispatcher were outdated and extremely impractical/difficult to oversee, making a modernization (which would also involve further safety-measures like accoustic warnings) overdue. The dispatcher’s above-average workload further made things difficult, leading to the recommendation to double-staff the position with two dispatchers. It was also recommended to reevaluate the “human factor” and pay more attention to it during the planning of refurbishments and/or new signal boxes as well as during training, recommending an emphasis on training for abnormal operations. Lastly, it was recommended to teach train drivers about the essentially identical characteristics of the adjacent rail line between Penzing and Hütteldorf, whose routing was seen as sufficiently similar to divert trains onto that section even if the drivers weren’t certified for the entire length of that line.

4024 101’s leading car, showing severe but still considerably less damage.

4024 101 was rebuilt after the accident, combining reusable cars from both trains. One unneeded middle car was used to expand a three-car 4023 to a four-car 4024. The damaged end-cars were placed in storage at a maintenance facility in Vienna, likely being used for parts before being scrapped.

The “new” ÖBB 4024 101, rebuilt from both involved units, photographed departing Penzing station for Hütteldorf in August 2019.

_______________________________________________________________

I have chosen to stay off Medium’s monetization offers to keep these stories as accessible as possible, but you can support me with a small tip via “Buy me a coffee” if you feel like it.

_______________________________________________________________

A kind reader is posting the installments on reddit for me, I cannot interact with you there but I will read the feedback and corrections. You can find the post right here.

--

--

Max S

Train crash reports and analysis, published weekly.