Downhill Disaster: The 2006 Dürrenast Runaway Train Collision


Dürrenast is a District in the south of Thun (pronounced more like “Tuun”, there is no “Th” in German), a city of 44271 people (as of 2019) in central Switzerland. Thun is located on the northern and northwest shores of Lake Thun (“Thunersee”), 27km/16.8mi south-southeast of Bern, the Swiss capitol, and 98.5km/61mi southwest of Zürich (both distances measured in linear distance).

The location of Dürrenast in Europe.

Dürrenast lies on the Thunerseebahn (“Lake Thun Railway Line”), a partially double-track electrified main line running from Thun southbound via Spiez to Bönigen. Opened in it’s full expansion in 1901 the line is owned today by the BLS AG, a railway company supplementing the SBB’s (Swiss national railway) services owned mainly by the Canton of Bern. While regional traffic was recently reduced there is still a lot of long distance passenger and freight trains using the railway line. Dürrenast had it’s own station on the line until 2005, since then it relies on nearby Thun main station just 1.2km/3938ft (linear distance) away for railroad connections.

The approximate site of the main wreckage. Note the two silver-roofed buildings to the north, they will show up again.

The train involved

The night of the accident the BLS was working on a set of points in Blausee-Mitholz, 25km/15.5mi (linear distance) from Dürrenast. The five man crew was working off Train number 36240, at the time a seven car train owned by the BLS and two construction companies. The train at the construction site consisted of a crane, the locomotive, two “protection cars” (two-axle flatbed cars with equipment on them), two MFS ballast cars and two AVES conveyor belt cars (used to move gravel into the ballast cars). The locomotive used was BLS Tm 235 099–9, a two-axle 38 metric ton “tractor” with a power output of 588kw made by Stadler. “Tractors” are small but relatively powerful locomotives used for shunting or maintenance trains. Number 99 had been delivered in 1996, and so far had had an uneventful service life.

BLS 235 099–9, the tractor involved in the accident, photographed in 2002.


On the 17th of May 2006 at approximately 0:40am the construction crew is finishing their work at Blausee-Mitholz, 25km/15.5mi (linear distance) south of Thun. Aside from the train they also use a road-rail excavator to move gravel into the ballast cars via the AVES-units. At the same time another 11 man construction crew is working in Dürrenast, removing remains of the retired station and installing noise protection walls.

The location of the construction site relative to Dürrenast.

At 0:45am the train departs the construction site headed for Frutingen (6.5km/4mi linear distance away), followed five minutes later by the excavator. The plan is to de-track the excavator there and have it’s driver (who’s a licensed train driver) help shunt the train. As they arrive at Frutingen they discover that one of the AVES-units had a faulty brake, leading to it travelling some of the way with a stopped axle. This caused a flat spot necessitating removal of the unit and it’s protection car from the train. During the lengthy procedure of rearranging the train the crane is also moved to the back of the train. Now five cars long, including two ballast cars full of old gravel headed to a concrete plant, the train has an approximate weight of 260 metric tons.

A protection car (left) and ballast car (right) identical to those in the train involved. These were placed right behind the locomotive.
An AVES-unit and small crane car, identical to those in the train involved. Located behind two ballast cars these were at the back of the train.

Tragically, during the shunting work, the brakes weren’t properly tested. As such, the closed pneumatic valve on the back of the locomotive wasn’t discovered. This led to the entire train except for the locomotive being left with inoperable brakes on a route that would go relatively steeply downhill. Who exactly closed and failed to reopen the brake valve could never be determined. As the unplanned shunting work overran the schedule two of the construction workers were told to go home, the two drivers would proceed to a concrete plant near the town of Hondrich with only one worker/machinist to help unload the two ballast cars. The plant was 9.5km/6mi away, with a largely downhill track going there at an angle of up to 15 Promille/1.5 percent.

The distance from Frutingen to the siding at the concrete plant.

The accident

The train left Frutingen at 3:01am. Different sources say either all 3 men rode on the locomotive, or the machinist remained in a container/brake room on the first car (it’s unknown what kind of container was on the protection car). At 3:05 the train driver operating the locomotive radioed the dispatch in Spiez saying he had next to no deceleration when he applied the brakes. Due to hectic and high noise aboard the train the dispatcher had to ask the driver to repeat his words several times, by the time the situation became clear the train was travelling at 80kph/50mph and was only 3km/1.9mi from the plant. Records aboard the train would later show a completed and successful brake test, creating a mystery as to why the brakes wouldn’t work. The driver asked for the train to not be directed into the siding, knowing it would either derail in the tight turns or slam into the buffer at the end of the siding at high speed. The dispatcher obliged, setting a path to keep the train on the main line. He did this under a shunting operation, leaving the signals red. He was hoping that the defect was between the controls and the brakes, so passing a red signal way outside the station would auto-stop the train. Shortly after passing the siding the train went into a tunnel, loosing contact with the dispatcher. With immediate disaster avoided the dispatcher at Spiez contacted his coworker at Thun, looking for a way to stop the runaway train. The official guidelines for runaway vehicles had clear instructions for a case like this. First, if possible, endangered trains and personnel are to be moved to safety/out of the way. Secondly, should there be no way to stop the runaway vehicle, an attempt to let it roll to a stop on level or uphill sections is to be made. Thirdly, if this is unsuccessful or impossible, a controlled derailment is to be initiated, sacrificing the runaway vehicle to protect other trains, houses and people. In any case, a solution has to be found that doesn’t increase the danger to people and property around the tracks.

  • A diversion towards Simmental was not chosen since the train would run into an active construction site there. The tracks had been removed, and the train was too close to evacuate workers in time. Moreover, the train would have to navigate several sets of points at Spiez station designed for 40kph/25mph at most, going over twice that, as well as several barrier-less level crossings.
  • Another diversion towards the lake at the town of Gwatt was also cancelled, due to the same 40kph switches being located there too, in a residential area.

By this point, 2.4km/1.5mi ahead of Gwatt, the train was already travelling at approximately 100kph/62mph. This was even faster than the locomotive was engineered to drive (75kph/47mph), showing the locomotive was being pushed by the train.

Entering Thun the train could be directed into two dead end sidings, but those were lined up with a building that was part of the area’s gas supply. Fearing a massive explosion this option was cancelled also. Unfortunately for everyone involved there had been an incident earlier in the day that had shut down railway traffic north of Thun, so as the runaway train approached the city Thun main station was packed with trains, there was no empty track to let it run through the city and onward to the north. This would’ve been the best outcome, as the track (which had gotten a little less steep behind Gwatt) would level out there. The least bad option would’ve been a “rolling road” train with semi trucks on it, but not knowing what those trucks were loaded with the dispatchers couldn’t risk a catastrophe should the runaway train strike a truck with dangerous cargo in downtown Thun, such as a tanker truck with fuel.

A rolling road with a tanker truck, this is what the dispatchers feared the train could hit.

Running out of time, track length and options the dispatchers accepted that there was no way to stop the train intact, and few ways to derail it. Their choice fell on the construction site at Dürrenast, two parked, unoccupied groups of freight cars located approximately 500m/1640ft apart. Two cars were parked at kilometer 1.4 (the distance is counted from Thun southbound, while the train moved northbound) with their brakes applied, and another 6 cars parked at kilometer 0.9.

The location of the first impact (orange) and the final location of the wreckage (red).

The “stacked” subsequent impacts were the best way to derail the train and stop it, if not the only way by this point. The dispatcher at Thun, now completely alone with the runaway train, set a 60kph set of points ahead of the construction site to lead the train into the right hand track (trains in Switzerland run on the left) and radioed the foreman at the construction site ordering him to tell his crew to immediately vacate the site as there were only minutes left. A worker tried to place a wheel chock, a device to help stop (or in this case derail) a train, on the tracks at the last moment, abandoning the attempt when he saw the headlights in the distance. The attempt would have been futile anyway, in the confusion he had started placing them on the usual, left hand track.

A wheel cock in use, keeping a parked car from rolling away.

At 3:20am the runaway train struck the first group of cars, which were full of scrap metal, at approximately 89kph/55mph. The dispatcher hadn’t known the cars were loaded, so the impact was a lot more violent than expected. The train pushed the two cars along to the second group, striking these (loaded in part with concrete segments) just seconds later. The impact telescoped the first two struck cars as the tractor and the first car derailed. The small tractor got crushed between the two trains as it was thrown off the tracks, with the wreckage stacking 3–4 layers high. All 3 people aboard the runaway train died during the second collision, while the 11 workers at Dürrenast escaped uninjured.

The wreckage as seen the next morning, the orange square used to be the cabin of the locomotive.


16 minutes after the driver had radioed dispatch with a brake-issue a deafening crash awakened residents in Thun. People close to the site of the accident looked out their windows to find a massive pile of torn and twisted metal where the railway line had been.

A photo from the report, indicating what part of the wreckage belonged to which train.

Local residents were the first people to reach the scene, soon followed by the fire department which had to extinguish a small fire that had started at the frame of the locomotive. It was soon clear that nothing could be done for the crew of the runaway train, initially their bodies couldn’t even be accessed. Responders worked throughout the day, cutting up the wreckage piece by piece to recover the bodies. The work took hours since they had to constantly ensure the stack of wrecked rolling stock wouldn’t collapse on them. The collisions had destroyed the tracks, making usage of the important line completely impossible.

An aerial view of the scene, the impact barely moved the parked cars. Also note the silver-roofed houses mentioned above for a sense of location.

With the bodies removed investigators soon focused their attention on the brakes of the train, the obvious cause of the collision. While the cars showed no sign of brakes being applied the brakes on the locomotive had been worn down entirely, and signs of excessive heat were found on the wheels. This discovery matched a witness at Spiez station, who saw the train pass with the brakes on the locomotive glowing red hot. It transpired that the train had traveled the entire way with the locomotives brakes more or less fully applied, burning away the pads, while the rest of the cars had rolled freely, accelerated by the downhill route.

The brakes of the locomotive being examined, showing signs of excessive heat caused by applied but insufficient brakes.

Collecting piece after piece from the train soon showed the cause of the brake failure, as the air valves on the back of the locomotive’s frame were found closed with nothing indicating that the collision closed them.

The handles for the valves on the back of the locomotive, closed with no indication of an impact during the collision.

Looking further the valves on the ballast cars were found to be in the “empty”-setting, a setup allowing lower pressure for the much lower weight of an empty ballast car. However, they were still operational and would have slowed the train, just over a longer distance. With the knowledge the brakes on the train were fine but just not hooked up the investigators went on to check how much braking-power the train actually had. In June 2006 investigators took an identical locomotive to a nearby section of track and simulated varying degrees of brake failure, seeing how the locomotive reacted, as well as doing a reenactment of the runaway train’s path between Frutingen and Spiez. Testing a comparable train of 295.5 metric tons on 15 Promille of descent (the steepest sections the runaway train went down) showed that a fully connected braking system could keep the train under control with no issue. Furthermore, the experiments found that the brakes on the locomotive alone could have stopped the train at up to 50kph/31mph. The BLS dictates that after departure a brake-test has to be done at no more than 30kph/18mph. During this time recordings also surfaced showing that both drivers had not fulfilled minimum break times, operating under a lack of sleep/rest.

Publishing the report in April 2007 the investigators concluded that, after rearranging the train at Frutingen (and running way behind schedule) only one of the drivers checked brakes of the train. He presumably found a fault with one of the cars and, instead of resetting the system and starting the stationary tests over he shut the valves on the locomotive, bled the system (releasing the air and thus releasing the brakes), and signed off on the system working, forgetting to open the valves again. Another indication to this were statements from the two workers sent home at Frutingen, who were allowed to head home as the tour from Frutingen to the concrete plant lengthened their way home and “it was getting really late”. In summary a sleep-deprived train driver aching to get the train to it’s destination and go home acted negligent when checking the brakes, setting the train up for disaster. Had the drivers then done the mandatory check at 30kph/18mph the fault would’ve been found and the train stopped safely, but both men seemingly neglected to do this check, dooming the train and themselves as well as the machinist traveling with them.

The report closes with another few notes, none of which played a role in the disaster happening or not. It says that the brake-line between the second and third car was not connected. However, this could have happened during the collision. Also, the locomotive was fitted only with a simplified data-logger recording speed on the last 3900m/2.4mi, a type not permitted for locomotives actually travelling on open track. So technically, the small locomotive should’ve never taken the train from Blausee to anywhere. Furthermore, the locomotive’s control desk operates the brake via one lever used for acceleration and deceleration. The inability to pick exclusively the pneumatic brakes or the engine’s dynamic brake (the latter cannot stop the train at higher speeds) means even if done properly the 30kph brake checks aren’t completely reliable in their results as, at that speed, the dynamic brake helps to stop the train.

With the people at fault for the events dying in the events the report closes on recommending that drivers are reminded more explicitly of the importance of proper procedures, including rest times, and that the data-logger on the locomotive is upgraded if it is to keep pulling trains on the open track.

A few weeks after the accident the damage from the collision is repaired, five years later the meadow the wreckage sprawled into is replaced by a new gym and it’s soccer field. The BLS still owns the second Tm 235 locomotive used during the investigation. With the train not being to blame for the events there was no need to retire it. In 2019 the SRF (Swiss television channel) produces a documentary on the accident, containing original footage, recorded radio calls and interviews along with reenacted scenes. While there is no memorial at the site the events certainly aren’t forgotten by the locals.

The SRF documentation on the events, I couldn’t find an English language version.


Join the discussion about this post on Reddit!




Train crash reports and analysis, published weekly.

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Five Ways to Learn About U.S. History

Escaping the Loma Prieta Earthquake

The Second Industrial Revolution

Captain Cook Coined the Name “Nootka Sound.” Literally

Hazarat Nizamuddin Dargah

Funding Bicycle Urbanism, Now.

Mohawk Resistance —Warrior and a Canadian military soldier.

Mell H. Embler Sr.: WWII Marine Corps Mechanic

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Max S

Max S

Train crash reports and analysis, published weekly.

More from Medium

Slippery Slope: The 2009 Kaštela (Croatia) Train Derailment

The Long Shadow of War: The story of the shootdown of Iran Air flight 655

Not Turning Back: The 2011 Münchenstein (Switzerland) Tram Derailment

The Captain’s Gambit: The crash of Allegheny Airlines flight 485