Escaped Express: The 2016 Andermatt (Switzerland) Runaway Train Derailment
Background
Andermatt is a municipality of 1588 people (as of 2023) in southern Switzerland, located in the canton of Uri 51km/32mi south of Lucerne and 52km/32mi east of Lauterbrunnen (both measurements in linear distance).
The municipality is the starting point of the Schöllenbahn, a 3.77km/2.3mi meter-gauge rack-railway connecting Andermatt with the Gotthard Railway at Göschenen. Meter-gauge is a narrow gauge standard commonly used in Switzerland, placing the rails 1000mm apart (instead of the standard 1435mm), while “rack-railway” refers to a toothed rail in the center of the track engaging with a cogwheel under the locomotive to help the train up or down particularly steep sections where the normal grip between rail and wheel would be insufficient for safe operation. Being a narrow-gauge railway with a rack-system allows the Schöllenbahn, which opened in 1917, to navigate tighter curves and steeper inclines than a regular railway. The alpine terrain still required over 2km/1.25mi of tunnels and galleries (open-sided tunnels for rockfall-protection) and 4 bridges. The line was originally primarily built for military purposes but is in civilian use nowadays, carrying around 400 thousand passengers per year. Operations are handled by the MGB (Matterhorn-Gotthard-Bahn/Matterhorn-Gotthard-Railway), a state-owned railway company formed in 2003 by merging 3 smaller companies.
The Vehicles Involved
The MGB used a small “tractor” to shunt cars around Andermatt station, which is a term used in Switzerland for a small, usually diesel-powered shunting locomotive. Duties were performed on the day of the accident by MGB Tm 2/2 number 4972. The locomotive was one of two made of the model, originally manufactured in 1961 for a German cement-factory’s on-property railway before being exported for and adapted to the MGB in 1976. The Tm 2/2 weighs 21.8 metric tons at 7.2m/24ft in length, with a 12 cylinder diesel-engine driving both axles.
The locomotive had been hooked up to a four-car train of passenger cars on the accident, made up of two pairs of a middle- and a control-car each. The two middle cars were BDk 2236 and 2237, each measuring 18.5m/61ft in length at 17.8 metric tons in weight, while the cab-cars (largely identical except for a driver’s cab to remotely control the locomotive) were ABt 4193 and 4194. The type BDk had been in service since 1986 while the ABt had been introduced in 1971. There were no people aboard the four cars as they were being moved around the station. Judging by photos from the aftermath they were coupled to the long-hood end of the locomotive.
The Accident
Four passenger cars are pushed into a parking position on the northeastern side of Andermatt station on the 1st of September 2017 at 6:42am. The handbrake on the easternmost passenger car (BDk 2236) is applied before the tractor is uncoupled from the cars and used for other shunting-work around the station for approximately 50 minutes. The track where the cars are sitting (referred to as “curve 4” for the curve it leads into) has a slight downhill angle to the east, but the handbrake can hold the train there just fine.
The tractor returns to the eastern end of the train at 7:40am, being coupled to it and having the pneumatic lines connected. The pneumatic shunting brake on the tractor is applied along with its handbrake before the passenger car’s handbrake is released. The shunting worker then abandons the train to help elsewhere.
A scheduled express service departs Andermatt station for Göschenen at 7:48am. The dispatcher spots the shunting train starting to move at 7:49am and radios the shunting worker who had parked it, reminding him that the train has no permission to move and has to be stopped immediately. The worker responds that there is nobody aboard before sprinting after the train, which departs the station and rolls through an unsecured level crossing. It is pure luck that no cars are struck as the train comes through the crossing with no warning. The dispatcher, in the meantime, radios the driver of the recently departed express train and informs him that his train may be caught up to by runaway rolling stock, leading the driver to instruct his passengers to abandon the rear cars in favor of the forward ones.
The worker, who has yet to get the runaway train back into sight, has almost reached the Bäzberg Tunnel when he hears a loud crash from the darkness. The train has dived into the tunnel approximately 320m/1050ft past the level crossing before reaching the following gallery after another 380m/1247ft. The S-turn during which the tunnel turns into the open-sided gallery proves too much for the train to handle under its ever-increasing speed, causing it to derail and grind along the final few meters of the tunnel wall before striking one of the gallery’s support-pillars. The impact stops the tractor dead in its tracks, obliterating most of the driver’s cab, and severely damages the following passenger car. The derailing train has also torn down the overhead catenary, stopping the scheduled train up ahead from lack of power, but nobody is physically harmed in the incident (the shunting worker suffers a shock).
Aftermath
Investigators take a statement from the shunting worker who parked the train and turn their attention towards the braking-systems of the tractor. Two braking-systems on board were used ahead of the accident, a “parking brake” which is mechanically actuated and uses spring-pressure to push brake pads against the wheels of the train and a pneumatic shunting brake which uses compressed air for the same purpose. According to the shunting worker both were set when he left the train. The investigation finds that the handbrake on the tractor had at one point been operated by a hand crank before being converted to the spring-loaded system, during which the grip of the brake was calculated at 11kN. A calculation of the parked train’s weight on the ever so slightly (0.4%) sloped track shows that 3.45kN would’ve been sufficient to hold the train in place.
Investigators recover the components of the tractor’s handbrake from the wreckage and examine the remains of the brake pads. Both pads show the expected wear-markings but also blueish discoloration in selected spots. Those marks are caused by excessive heat, which implies that the brake was at least partially applied as the train started rolling. That does confirm the shunting worker’s claim of having set it, but doesn’t explain how the train rolled away anyway. The condition of the tractor after the crash along with the additional damage resulting from some cutting required during recovery made a check of the system’s functionality impossible
The remaining Tm 2/2 tractor is examined by the investigators in order to check if the braking system is consistent with the supposed capabilities. They find that the new brake pads are measuring 60mm in thickness, too much to install in the handbrake-system. Workers explain that they developed a habit of using identical, used pads from the pneumatic shunting brake once those are worn down to a point where they fit. There is no official guideline for pad replacements on the handbrake, as the brake is only used when the train is stationary and thus isn’t exposed to wear.
The investigation turns its focus back to the blueish marks on the brake pads, figuring that a properly applied brake pad with insufficient pressure wouldn’t have heated up like it did in isolated spots but more evenly across the entire surface of the pad. The localized spots indicate that the pad was unevenly touching the train’s wheel, something the usage of the brake couldn’t even out as its use as a parking brake meant the wheel usually wasn’t spinning and thus couldn’t remove dirt and rust or fine tune the shape of the pad. The report explains that the spring-mechanism only moves the pads by 4.5mm and can’t be adjusted, meaning the pads were likely only getting insufficient contact with the wheel, generating a pressure in the neighborhood of 6kN instead of 11. The problem didn’t exist with the original hand-crank system, as that mechanism effectively pressed the pads against the wheel as far as the crank could be turned by whoever operated it.
But there was still the pneumatic brake, which in itself was meant to be able to hold the train. So even if the handbrake was faulty, why did the pneumatic brake fail to keep the train in place? As it very soon turned out: Because it was broken in at least one way. Investigators recovered the main brake valve from the wreckage and set it to the “full stop” position, only to see the handle snap out of the setting in several attempts. There were two of these valves on the tractor, and both turned out to be faulty. The left hand one (used ahead of the runaway) would not stay in the “stop” position, while the ratchet-mechanism keeping the handle there on the right hand valve was very easily overcome. Disassembly of both valves showed intact springs and actuators, but the gate for the “stop”-position was severely worn on both, especially on the left-hand one.
The investigation once more went to the identical tractor they’d previously used to compare the handbrake, and found that on it, too, the handle of the shunting brake valve wouldn’t stay put. This alone didn’t disable the brake, but it meant that the reservoir was no longer being refilled by the compressor. This allowed small leakages that tend to appear in the system (the reason why the compressor is usually kept “on alert” to refill the system if necessary) to gradually drain the pneumatic pressure from the train until the brake disengaged after two to three minutes. The investigators concluded that the worn mechanism on the crashed tractor caused the handle to leave the “stop”-setting 3–4 minutes after the train was parked, leading to a loss of air pressure that allowed the brake to disengage after approximately 8 minutes. The handbrake, which is the main way to keep a train secured when parked, provided insufficient grip (and thus braking-power), allowing the train to start motion at that point, slowly rolling away as the handbrake alone wasn’t able to keep the train stationary in its condition.
The MGB introduced a temporary rule that demanded that the “surviving” tractor be secured with a brake shoe as a third measure, a practice that ended when its handbrake-system was modified to also act on the brake pads of the shunting brake. This ensures that sufficient braking-force is provided regardless of the condition of the handbrake’s exclusive two pads. The investigation further recommended that maintenance-guidelines for the handbrake-system be introduced, ensuring specifically the adjustment and braking-force of the handbrake.
ABt 4194 and BDk 2236, the forward pair of passenger cars, were written off after the accident and scrapped, while the other two cars only suffered minor damage and were repaired.
_______________________________________________________________
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.
