Precarious Points: The 2020 Lahnstein (Germany) Train Derailment
Background
Lahnstein is a city of 18719 people (as of 2024) in western Germany, located in the federal state of Rhineland-Palatine 58km/36mi southeast of Bonn and 79km/49mi west of Frankfurt/Main (both measurements in linear distance).
Lahnstein lies on the Lahntalbahn (Lahn Valley Railway), a 103km/64mi double-tracked partially electrified main line connecting Wetzlar (Hesse) with Koblenz (Rhineland-Palatine). The line was opened in sections between 1858 and 1863, becoming mainly a freight corridor especially for military transports (gaining it the name “Kanonenbahn”, “Cannon Railway”). It now sees everything from regional trains to long distance freight services, being limited to no more than 120kph/74mph due to its many curves and inclines as it follows the Lahn river.
The Train Involved
DGS 49077 was a chartered freight service from Rotterdam (Netherlands) to Basel (Switzerland), consisting of 19 four-axle “Zans”-type tank cars carrying diesel fuel. Both length and empty weight of the tank cars varied slightly, with the cars measuring around 17m/56ft in length (varying by a few centimeters) at a weight of 23.6–24.3 metric tons. Each tank car was carrying around 63 metric tons of diesel fuel. They have a top speed of 100kph/62mph. All tank cars had passed recent checks and re-certifications to haul diesel fuel. The tank-cars were rented from different rolling stock providers like GATX and Ermewa.
Pulling the train was BLS Cargo 193 496, a multi-system multi-purpose electric locomotive introduced by Siemens in 2010. The unit involved in the accident had been manufactured in 2018 and was equipped to run in the Netherlands, Germany, Austria, Switzerland and Italy, allowing it to haul trains across large parts of Europe without needing to be swapped for another locomotive along the way. The locomotive measured 18.98m/62ft in length at a weight of 90 metric tons and had a top speed of 160kph/99mph, although DGS 49077 wasn’t allowed to go faster than 90kph/62mph.
The Accident
DGS 49077 is approaching Niederlahnstein station at approximately 6:25pm on the 30th of August 2020, just over an hour after the driver took over from a colleague at Cologne-Ehrenfeld station. The train passes the entry-signal for the station at 55kph/34mph, 5kph/3mph below the local limit, before the driver slightly accelerates. Normally the train control system should’ve ensured a 55kph/34mph speed limit by the time the train passes the following section-signal (Zsig), instead the train passes it at 70kph/43mph at 6:33pm. The path set by the dispatcher now requires the train to change from one track to the next one to the right at the points number 35. The driver first activates the electro-dynamic brake of the locomotive (effectively decelerating by using the motors as generators) before adding input via the pneumatic brake.
Within a moment the driver feels a light impact on the rear, triggering an emergency stop. Behind him, the leading tanker car has run into the locomotive on the left hand side and levered its own leading axle off the rails, destroying the points number 18 and 17. The leading tank car proceeds to fall over due to pressure from the 17 tank cars behind it and as it now lacks the guidance from the rails, pulling the following cars with it. Their mass completely destroys the tracks while impacts from the buffers (mushroom-shaped impact elements on the end of a rail vehicle) with the tanks itself cause punctures, allowing diesel fuel to leak out. The leading car lifted the locomotive’s rear axle out of the rails as it fell over, tearing off the coupler and severing the pneumatic brake lines. Brakes now fully apply throughout the train, bringing the locomotive to a halt a short distance down the track partially derailed but upright and mostly undamaged at 6:35pm. Behind it eight tank cars have derailed, six of which fell over and leak cargo, with the leading two cars leaking the most due to large gashes in the tanks. The official report lists the driver as “slightly injured”, likely referring to a shock.
Aftermath
The derailment happened pretty much right in front of the station’s dispatcher, who proceeded to lock down all tracks through the station and alert emergency services. The first responders on site were firefighters from a near-neighboring fire station, who arrived within 4 minutes of the alert. Several local residents were evacuated as a “ban zone” was established until the fire risk could be properly estimated and controlled. Later estimates show that around 180 thousand liters/47.6 thousand gallons had leaked from the train cars.
Investigators from the BEU (Bundesstelle für Eisenbahnunfalluntersuchung/Federal Authority for Railway Accident Investigation) find derailment-marks on the left-hand rail in a right-left S-turn between points 35 and 18, indicating that at least one axle of the leading tank car derailed at that point and did so to the inside of a turn. The barely derailed wheel then struck the “wing rail” (a wheel-guide) on points 18, being forced further out of alignment. A much further reconstruction of the derailment wasn’t possible due to destruction of the track, which ends at the damaged tongue (the part of the points which moves) of points 18 right behind the struck wing rail.
The investigation, curious about how a train managed to derail to the inside of a turn, closer examined the design and construction of the area around points 18 and 35. In the direction of the train (from 35 to 18) the track first formed a right-hand turn with a radius of 350m/1148ft, followed by a 14m/46ft straight section and finally a left-hand turn into points 18 with a radius of 150m. Both points were designed for a 40kph/25mph speed limit when used in a “turning” setting rather than “straight ahead”. The radii of 150m/492ft was tight enough to require special approval, only to be given when there was no wider radius possible. However, even with special permission given the speed limit for a 150m/492ftft radius was 30kph/19mph at most, not 40kph/25mph. Points 35 had been replaced in 2010, and documentation for the calculation that gave the excessive speed limit couldn’t be provided by DB Netz (the company in charge of German rail infrastructure) when requested by the investigation. In simple terms, there was no evidence of when the 150m/492ftft radius had been installed and how the speed limit for it had been approved.
What was provided to the investigation was paperwork from when points 18 were replaced in 1996, showing a radius of 190m/623ft for the now tighter turn. Why the radius wasn’t built to that spec or why the 40kph/25mph speed limit was maintained couldn’t be explained. A track-plan from 2008 showing the tighter radius and a 25kph/16mph speed limit didn’t help, especially since there was documentation for several adjustment-procedures to make the lower speed limit reality, all of which were incomplete. There were evidently plans to replace large amounts of track surrounding the site of the accident within 2020, bringing back a wider radius, but that project didn’t come to fruition due to the derailment. The DB Netz did continuously argue that operation of the tighter radius was allowed under a special section of the German Railway Construction and Operation Requirements (EBO), but this law was meant for temporary situations until a proper radius could be accommodated and still didn’t explain why the excessive speed limit wasn’t addressed.
The problematic design of the section wasn’t unknown, in fact a supervising driver watching as a trainee drove a train through the station reported a “notable jolt” 3 days before the accident when passing the site of the accident. He was told by the dispatcher that there was a known issue with the track in that spot and that there had already been work done on the spot half a year ago (which apparently hadn’t had any effect on the problem). The dispatcher was required to record the reported issues with the track in the day’s paperwork and also should’ve locked the section down until it could be examined, but he did neither.
That situation alone was bad enough, but even at 40kph/25mph the train should’ve made it through the site. After all, trains regularly did. The problem was that the train wasn’t going at 40kph/25mph, it was going around 60kph/37mph. Investigators thus looked into the recordings from the train’s data-logger (similar to an airplane’s “black box”), and found problems literally from the start of the driver’s shift. After taking over the train at Cologne-Ehrenfeld station he ensured that the train’s formation matched the paperwork and proceeded to (wrongly) calculate the braking-performance of the train. He then entered the resulting data into the PZB-system (Punktuelle Zugbeeinflussung/Punctual Train Control), along with a scheduled top speed of 100kph/62mph. The actual scheduled top speed was 90kph/56mph. The wrong speed and faulty brake-data resulted in the system classifying the train as a category M-train when it was actually a U-train with a higher weight/lower brake-performance, resulting in different speed-limitations. PZB-classifications go from “O” (“upper”) to “M” (“Middle”) to U (“lower”), accommodating different trains’ different behaviors especially regarding braking capability. The false classification meant that the system was more “lenient”, as it would expect the train to be able to decelerate faster than it actually could.
The train departed Cologne-Ehrenfeld station at 5:10pm on a 116 minute delay, with the driver soon breaking the 90kph/56mph speed limit his train had by going as fast as 107kph/66mph. A category M-train is allowed up to 125kph/78mph, had the PZB-system been properly programmed it would’ve kicked in when the driver attempted to pass 105kph/65mph. The train eventually reached the entrance-signal to Niederlahnstein station, passing it at 55kph/34mph. The PZB-system, still working with data for a category M-train, ensured that the train wouldn’t go past 70kph/43mph, when it should’ve guarded a 40kph/25mph-limit. This allowed the driver to pass the section-signal at 61kph/38mph and actually increase his speed to 62kph/38.5mph before reaching points 35.
The driver started decelerating at 6:35:11pm, adding the pneumatic brake to the electro-dynamic one a second later. Another 5 seconds and 85m/279ft later the driver, likely feeling the impact from the train on the locomotive as a light jolt, triggers an emergency stop at 57kph/35mph. The train dumps air pressure, the brakes apply throughout the train, but the damage was done.
The leading tank car ran into the back of the locomotive 2 seconds after the emergency stop was initiated, marked by a jolt on the data-logger. The unusually tight turn along with the deceleration at the same time as the impact occurred caused the tank car’s forward left buffer to become lodged under the locomotive’s rear left buffer after sliding off the buffer plate, briefly lifting the locomotive’s rear axle out of the rails before the locomotive pushing down on the derailed tank car’s buffer produced enough force to tip the tank car over at 49kph/30mph. Damage between the locomotive’s rear buffers shows where the tank car likely collided once more with the locomotive’s frame in the process, proving that the buffers were past each other. The tipping-motion tore the locomotive off the train, leaving it to come to a stop by itself 82m/269ft further down the tracks at 6:35:33pm.
The report notes that two of the tank cars (car 2 and 7) suffered holes in the tank wall caused by colliding with adjacent cars above the buffer-line, leading to a total loss of cargo on those two cars. Some tank cars do have additional impact-elements to protect against such collisions, but those are not required for the cargo carried by the train and weren’t installed on the involved tank cars. The report found that the tank cars were likely fully intact before the accident, and that the excessive speed paired with the unusually tight turn caused cargo to slosh from side to side, aiding in the weight-shift that caused the train cars to derail and (in some cases) fall over.
The investigation proceeded to look further into the driver’s professional history, as his false operation of the PZB-system and voluntarily excessive speed were seen as key factors in the cause of the accident. They found that he was working part-time for more than one rail service provider, and that two providers had stopped working with him and revoked their company-internal permits for him after incidents where he had disregarded speed limits. One of these incidents had been significant enough to require passing a training seminar afterwards, which the driver passed. Further investigation showed a different issue, as the driver had evidently posted on social media while driving trains, including videos of himself driving trains. This obviously clashed with a ban on private use of electronic devices while working.
Working as a train driver is marked by high responsibility and the ability to work alone with high concentration at various times of day/night, this required “exceptional trustworthiness” was seen as unfulfilled by the driver after the investigation, leading to the federal railway authority (EBA) revoking his license as a train driver in October 2020. The driver chose to pursue legal action against the decision and lost in November 2020. The court agreed with the EBA that wrong programming of the PZB-system and several speed limit breaches (even if the driver argue that they had been “brief and slight”) showed a failure to present the required exceptional trustworthiness required of a train driver, especially one entrusted with hazardous cargo. The public prosecutor’s office ceased the criminal investigation, explaining that the other factors contributing to the accident meant that the driver hadn’t acted negligent to a criminal degree. The excessive speed alone, enabled by the false data entered into the locomotive’s systems, wouldn’t have caused the derailment had the track been properly constructed/speed-limited, which in itself wouldn’t have caused the derailment (even with the excessive speed) had the report about the problematic track-position been properly forwarded since that would’ve seen the section locked down, keeping the accident-train off it.
The repair-effort involved the replacement of contaminated soil along with 19 thousand metric tons of gravel, 2km/1.25mi of cables, 300m/984ft of railway track and four points being replaced. The section of track destroyed by the derailing train wasn’t repaired and reopened until September 2020. Total damage to railway equipment, rolling stock and the environment was estimated at almost 20 million Euros/23.5 million USD. As of Summer 2025 there was still “significant contamination” in the soil, which was explained with not all the contaminated soil being removed as that would destabilize the infrastructure above. Instead a “bioventing”-process is being used since 2022, pumping oxygen into the contaminated soil so microorganisms break down the diesel. After 3 years of that process 70% of the contamination were gone, according to the German railway. A demand by local residents to restrict trains with hazardous cargo to 50kph/31mph or ban them from the line completely was rejected.
_______________________________________________________________
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. I will post the link as soon as I can.
