Background
Linz is a city of 211944 people (as of January 2024) in northern Austria, located in the state of Upper Austria 108km/67mi east-northeast of Salzburg and 30km/18mi south of the Czech border at Bad Leonfelden (both measurements in linear distance).
The Linzer Verbindungsbahn (Linz Connection Railway) is an urban rail line in the center of Linz, whose roots go back to a horse-drawn rail line which started operation in 1832. It has since been expanded and rerouted several times, now connecting Linz-Urfahr station in the north with Linz’ shunting yard in the south on 9.4km/5.8mi of track with trains going at no more than 40kph/25mph. The line sees no scheduled passenger services, but is used to shuttle trains between stations or to the local ÖBB (Austrian National Railways) maintenance facility. The shunting yard itself is located immediately adjacent to Linz’ harbor and a considerable industrial area which allows direct transfer of products from ships or storage facilities to trains for inland transport.
The Trains Involved
Z61004 was a freight train consisting of 15 six-axle hopper cars carrying iron ore. The train measured 244m/800ft in length at a weight of 1653 metric tons and was operated by CargoServ, a private Austrian rail service company who also provided the locomotive and driver. The train was pulled by CargoServ 193 267, a Siemens Vectron AC owned by European Locomotive Leasing and leased to CargoServ right after delivery in 2016. The Siemens Vectron is a four-axle multipurpose locomotive succeeding both the EuroSprinter electric locomotives and the EuroRunner diesel locomotives as it can be had with either propulsion. The configuration used by CargoServ is that of an electric locomotive which measures 18.98m/62ft in length at a weight of 86 metric tons. The Vectron-family has quickly become a phenomenon in the European railway market and can be found all over Europe in various roles for various companies. The freight train only carried the locomotive’s driver at the time of the accident.
Z 97209 was an acceptance and approval run by Westbahn, a private Austrian passenger rail service provider. The Westbahn had been in the process of introducing the class 4110 into their fleet, the second generation Stadler KISS electric multiple unit (EMU). Stadler has a history of “convenient” abbreviations when naming their trains, with KISS standing for “Komfortabler Innovativer Spurtstarker S-Bahn-Zug” (“comfortable, innovative, sprint-capable suburban train”). The Stadler KISS is a two- to eight-car bilevel EMU designed for express passenger services, introduced in 2011 with the KISS 1. The Westbahn used exclusively KISS Mk1 and Mk2 trains at the time, with the then-new Class 4110 describing 9 four-car units introduced throughout 2017. The four-car train measures 102.24m/335ft in length at a weight of 229 metric tons empty, and can carry 323 seated passengers in a two-class configuration at up to 200kph/124mph. The nine new trains were part of a package that also included an additional six-car train, costing the Westbahn 180 Million Euros/185.6 Million USD for the 10 trains. At the time of the accident a driver and seven passengers were aboard the brand new train, with six Stadler-employees and a Westbahn-employee overseeing the test-run before the train would be officially accepted by Westbahn.
The Accident
The brand new Westbahn KISS 2 is boarded by its driver and seven passengers at Wien Westbahnhof (Vienna West Station) at 8:20am on the 23rd of August 2017. The seven passengers are there to examine the train and its behavior during a test drive to Linz’ harbor, where the train is to park at Voestalpine, a local steel plant. The trip, which includes a top speed run, is meant to ensure that the new train was properly constructed and fulfills the requirements to be officially accepted by Westbahn ahead of starting commercial service. After some on-site examinations and last minute route changes the train departs Vienna at 10:14am. In the meantime, Z 61004 (referred to as the freight train from here on) has departed the fittingly named town of Eisenerz (iron ore) at 8:02am and is headed north, being intended to also eventually reach Voestalpine’s facility.
An uneventful trip sees the KISS arrive at Linz’ shunting yard by 11:36am, where the train stops at signal G8 after arriving on the oncoming (left hand) track. The local dispatcher radios the train driver as the train is stopped and tells them that they can’t follow the original route (labelled M 4037 in the report) but instead will be given a new route, “Patterm 47005”. The driver checks the company’s database on a company-provided iPad but can’t find the new route, as the database isn’t arranged by “patterns” but by route-numbers. He reports this to the dispatcher, who can’t tell him the route number but assures him that he’ll have a clear path to the parking position at Voestalpine.
Signal G8 switched to “clear” at 11:44am with an additional “attention”-marker, which the driver confirmed with the attention-button in his cab. The train next stops at signal Sch391 (Schutzsignal 391, “protection signal 391”), standing still for two minutes before the signal allows him to pass at 11:48am. The train proceeds across points 200 and 201 into the right hand track, turns a left hand corner and is faced with the oncoming freight train which is just approaching points 204 where it is meant to be directed off the KISS’ track to the left.
The KISS’ data-logger registers an emergency stop being triggered at 11:48:56am, 25m/82ft ahead of Signal Sch396 as the train is travelling at 37kph/23mph. Air pressure is dumped 3 seconds later, fully activating the brakes. However, the train runs out of track before it can come to a stop, leaving it to pass the points 14m/46ft behind the signal and run into the forward right hand corner of the CargoServ locomotive, which is itself travelling at 33kph/20.5mph, at 11:49:02am. The KISS “hooks” itself on the right hand buffer of the locomotive at 22kph/13.7mph, suddenly introducing extreme yawing-forces to its crash protection structure. The entire driver’s cab twists to the right and subsequently tears off the left hand mounts on the train car’s frame, getting pressed into the side of the freight train as the Vectron is derailed to the left. The crash protection structure (somewhat comparable to a race car’s roll cage) built into the KISS’ cab mostly withstands the forces pushing on it from opposite sites, retaining sufficient survival space for the driver as the mostly detached cab is dragged along the freight train’s side.
The impact throws two standing passengers in the KISS to the ground before the trains come to a stop a few meters past the point of impact, while the other five passengers happened to be seated at the time of the accident. The freight train has remained mostly on the track it was meant to enter while the KISS’ leading car is now blocking two tracks and has also torn the overhead catenary. Its driver suffers severe injuries while four of his passengers and the freight train’s driver suffer minor injuries.
Aftermath
The freight train’s driver notifies the ÖBB’s emergency coordination center at 11:52am, who in turn notify emergency services at 11:55am and begin powering down the overhead catenary. The first responders reach the site of the accident just after noon, finding a derailed freight train with its locomotive sitting just past what can best be described as a beheaded passenger train. The cab of the KISS has torn off its train just ahead of the forward doors, creating a wide gash almost all the way around the train car’s circumference as the cab was forced to the right while the train derailed to the left.
The entrance area of the KISS’ leading car is filled with debris, but despite its deformation responders find the driver’s cab to have structurally remained largely intact, allowing them to quickly rescue the driver once they cut their way through the outside wall of the cab. The remaining passengers of the KISS were in the passenger-area behind the doors, which remained intact and allows access/rescue through the rear doors of the train car.
The KISS’ driver is finally able to give his statement to the investigation a few days after the accident, explaining that once he received permission to depart from signal SCH391 he understood the instructions as being meant to pull past points 204 and then reverse over them in order to leave the track he had come in on and reach intersection KR21 to the southeast.He doesn’t recall seeing any signal ahead of points 204, certainly not a red one, which may have been subconsciously aided by the dispatcher talking about “giving him a path” to the parking position.
So, in essence, that was the cause of the crash. The KISS’ driver ran a red signal due to paying insufficient attention to his surroundings, perhaps having lost some confidence from the last-minute switch to an unfamiliar route, and ended up crashing into the oncoming freight train whose driver was doing exactly as he was told. But, in simplified terms, crashing a train isn’t meant to be that easy. Modern rail traffic builds on the principle to always have two independent measures ensuring safe operations. If the driver makes a mistake (or purposely ignores a signal) another measure is meant to override his input. This other measure used to be a second driver in the cab, but nowadays most railways use computerization for that measure in order to be secured against human error or interference by hierarchy (an assistant not daring to question his superior).
The ÖBB used PZB90 (“Punktuelle Zugbeeinflussung 90”/”intermittent train protection 90") to control trains independent of human interference at the site of the accident. The system works via an electromagnet next to the rails at each signal, matching a magnet on the train. A pre-signal is equipped with a 1000Hz magnet, a 500Hz magnet is found at slow-speed zones and ahead of main signals and a 2000Hz magnet can be found at main signals and inhibition signals. SCH391 is a main signal and was thus fitted with a 2000Hz magnet. The magnets are energized when their signal/instruction is activated (speed limit in use, main signal turned red), allowing a magnet (and computer) on the train to pick up the signal. A yellow signal (“attention”/”slow”) still allows a train to pass over it, but the driver has to acknowledge via a button in the cab. The control system has the capability to bring the train to a complete stop if the driver fails to acknowledge the warning, even if a malicious driver would continuously apply full throttle.
Sch396 was red at the time of the accident, meaning its 2000Hz-magnet was energized. A train passing over an active 2000Hz-magnet is always brought to an immediate emergency stop, since the driver has either willfully disregarded a red signal, failed to spot it or become incapacitated in some way. If they were aware of it (the purpose of the acknowledge-button) they wouldn’t reach the magnet. The 2000Hz-signal can be overridden in selected irregular cases (such as a defective signal), but that procedure doesn’t apply in this case. And indeed, the train’s data-logger recorded a 2000Hz-signal being picked up at 11:48:59, just as the train was dumping air pressure. This happened 3 seconds after the throttle lever was abruptly moved from 70% forward to 100% reverse upon the driver seeing the oncoming freight train in his path. It’s not quite clear if he triggered the emergency stop or if the PZB-input took care of that at effectively the same moment. But if the signaling system and the brakes worked as intended, why did the train still crash into the freight train?
The investigation zeroed in on that part of the cause rather quickly. The magnet on SCH396 is just 14m/46ft from the points 204. Trains are allowed to approach the signal at 40kph/25mph when it’s green, with the KISS involved in the accident requiring 43m/141ft to stop from 37kph/23mph it was travelling at when it reached the signal. That’s more than twice the amount of track it had left. Guidelines at the time didn’t require a “protective distance” beyond signals reached at 40kph or less, meaning the short distance was completely legal to engineer into the placement of signals and points. The on-site evaluation also revealed that SCH396 had a tendency to “blur” with the catenary support pole behind it if a driver looked at it from the previous signal (SCH391), but visibility was deemed to be sufficient as nothing was obstructing the signal itself. The investigation does note that having the train “inch forward” from SCH391 to SCH396 wasn’t necessary, but was a regular habit in order to clear sections of the line as quickly as possible by keeping trains as far forward in a section as possible.
The investigation also notes that the conversation between the dispatcher and the driver may have been a minor contributing factor, with the dispatcher saying “I’ll give you a plan, plan 47005” (initially misremembered by the driver as “pattern”) and not being able to provide the route-number, instead asking the driver to trust him that he “has a path” to the parking position. Usually dispatchers are meant to use very specific wording, listing origin, destination, plan, route-book and route-number, making it a very formal order which the driver noted was not given. However, the driver also noted that this specific wording was very rarely used. So the dispatchers “relaxed” demeanor and choice of words was far from an unusual occurence.
The report goes on to list five minor collisions similar to the one at Linz which happened between March 2017 and July 2019, all of which happened due to overrun signals with some also noting poorly placed PZB-magnets as a contributing factor. The investigation concluded that the main cause of the accident was the accidental running of red signal by the KISS, which happened to be constructed with insufficient space beyond it to stop if there was oncoming traffic at the same time (the KISS couldn’t have stopped without sliding past the points). The signal was 10 years old by the time of the accident, not a time span that led to significantly heaver trains or stronger acceleration capabilities. The placement was up to regulation, which happened to be flawed. It just happened to be that, so far, there had been no reason to question the regulations about safety-distances beyond 2000Hz-magnets since their placement had always been a contributing/minor factor at most (and the accidents, luckily, had been relatively minor). The fact that the signal and the catenary support pole behind it were the same color was listed as a contributing factor while the report also noted a systematic factor in the sparse and insufficient communication between the driver and the dispatcher, which led to the driver misunderstanding the instruction (but also not inquiring about what the dispatcher meant).
An additional risk-analysis was performed on the signal-placement in October 2017, with the evaluating expert recommending that SCH396 be moved back 36m in order to achieve a distance of 50m between the signal and points 204. This would leave sufficient stopping distance if a train like the KISS ran the signal to keep it from entering the points. The Westbahn also announced improving their training for drivers, specifically around the complex and sprawling shunting yards at Linz and Salzburg and to include rarely used routes more frequently.
4110 610, the KISS involved in the accident is repaired after the accident, while no details are public it’s assumed that the Westbahn had Stadler provide a whole new end car due to the extend of the damage suffered during the accident. It’s unknown if the repairs were finished ahead of the Westbahn retiring their entire 17-train class 4110 fleet in late 2019 in favor of 15 third generation KISS-trains. The trains were sold to the DB (German national railway) and received a new paintjob and interior at Stadler’s production facility. Maintenance will continue to be provided by the ÖBB at Vienna (Austria). The DB uses the trains as part of their modernized Intercity-network (IC2).
_______________________________________________________________
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.
_______________________________________________________________
Thank you for sticking with me through another year and welcome to 2025! Going from weekly to monthly releases allowed me to keep this blog going in a regular fashion despite having less time available for it. Some posts were literally finished the day before they were published. If you need something to “fill the time” I can point you to Admiral Cloudberg, who covers aviation accidents here on medium in excellent, lengthy articles.
