Harmful Crossing: The 1962 Harmelen (Netherlands) Train Collision
Background
Harmelen is a town of 8155 people (as of 2022) in the center of the Netherlands, located in the Province of Utrecht 30km/18.5mi south of Amsterdam and 39km/24mi northeast of Rotterdam (both measurements in linear distance).
Harmelen lost its train station in 1936, but is still the site of two rail lines merging. The town is passed by the Utrecht-Rotterdam rail line, a 56km/35mi double-tracked electrified main line opening in sections between 1855 and 1858 before receiving today’s routing by 1953. As it connects two of the Netherlands’ main cities it quickly became one of the country’s most frequented rail lines, seeing mostly regional and intercity passenger services at speeds of up to 140kph/87mph.
Just north of Harmelen the Harmelen-Breukelen rail line splits off from the Utrecht-Rotterdam line. The 8.3km/5.2mi double-tracked main line to Breukelen opened in 1869 and today allows speeds of up to 130kph/81mph. Nowadays it only plays a minor role for regional passenger trains, being mostly used for freight trains seeking a direct route from Rotterdam’s large harbor-facilities to the northern and eastern Netherlands without needing to change direction.
At the time of the accident the intersection of the two lines was arranged in the form of a same-elevation crossover, which meant that trains heading from Rotterdam to Amsterdam had to cross the oncoming westbound tracks before continuing on their way. This maneuver was usually secured by means of a pair of signals ahead of the intersection. If a path crossing the oncoming tracks was set westbound trains (those headed to Rotterdam) would receive a pre-signal instructing them to “expect stop” and slow down, followed by a red signal which would see them stop ahead of the intersection. The same applied for trains towards Amsterdam if a train from Utrecht had priority.
The trains involved
Train number 464 was a regional service from Rotterdam to Amsterdam, provided by a NS (Dutch national railway) Materieel ’46. The Materieel ’46 (also referred to as the Mat ’46) is a two- or four-car electric multiple unit introduced in 1949 after being ordered in 1946 (hence the name). The train involved in the accident consisted of a four-car unit (number 700) towing a two-car unit (number 297). A two-car unit measured 44.7m/147ft in length at a weight of 89 metric tons empty, the four-car version measured 87.1m/286ft at 169 metric tons. Both versions could reach 125kph/77.7mph, with the two-car units carrying 125 passengers (four-car version: 269) in a two-class configuration. The four-car train also included an onboard restaurant, which is included in the passenger capacity. At the time of the accident the train carried a driver and an estimated 180 passengers.
Coming the other way on its trip from Leeuwarden to Rotterdam was express train 164. The 11-car express train consisted mostly (7 out of 11) of “Plan E”-Type passenger cars, a four-axle passenger car introduced in 1954. Each Plan E-type car measures 23.05m/76ft in length and can carry up to 72 passengers.
The leading, seventh and rear two passenger cars were older “Mat ‘24”-type passenger cars, former electric multiple units converted to passenger cars by removal of the propulsion system and pantographs. The type was introduced as a multiple unit in 1924 before being converted to unpowered cars in the 1950s. The cars included in the express train were former middle-cars, seating approximately 70 passengers (sources vary between 64 and 88 seats) per car.
Pulling the express train on the day of the accident was NS 1131, a Class 1100 four-axle electric multipurpose locomotive introduced in 1950. Each Class 1100 locomotive measures 14.11m/46.3ft in length at a weight of 83 metric tons. The type has a top-speed of 135kph/84mph thanks to a power-output of 2030kW/2720hp.
The accident
On the 8th of January 1962 express train 164 is travelling from Leeuwarden to Rotterdam, having collected a delay of six minutes as it leaves Utrecht station, 10.5km/6.5mi linear distance east of the crossover, in thick fog with approximately 900 people on board. Normally the express train would pass the crossover at 9:14am, the electric multiple unit (referred to as the EMU from here on for simplicity) at 9:18am and lastly another train would follow the express train’s path by 9:21. This pattern was repeating every hour.
Due to the delay the express train doesn’t contact the signal box for the crossover on schedule, leading the signal box worker to give priority to the EMU coming the other way, intending to head north to Amsterdam. The EMU passes the main signal west of the crossover at 9:18am, travelling at 75kph/46.6mph. The EMU is going past a green signal as it approaches the crossover, with the express train being expected to hold east of the crossover. Fatally, this won’t be.
The express train passes the pre-signal showing yellow (“expect stop”) at 125kph/77.7mph, just as the EMU enters the oncoming track to transfer to the rail line towards Breukelen. At the time there is no train control system integrated into the signaling-system, obeying speed limits and red signals is the responsibility of the train drivers alone. The express train’s driver triggers an emergency stop at 9:19am, presumably spotting the red signal racing towards him. It’s too little too late, the oncoming EMU is only a few hundred meters away. The EMU’s driver likely never saw the express train coming before it slammed into his train head-on, still moving at 107kph/66.5mph.
The impact throws the express train’s locomotive off the tracks to the right, crushing the driver’s cab and turning the locomotive upside down as it tears off the train. The leading passenger car mounts the frame of the EMU’s leading car, causing the latter to sleeve over the passenger car and obliterate the EMU’s leading car’s interior. The EMU’s second car is torn open down one side as the express train’s leading five cars deflect off the track after passing through the remains of the EMU’s leading car, with the EMU’s third car being ground to pieces between the derailing express train’s cars and the rear three cars of its own train. Only the express train’s rear six cars and the rear 3 cars of the EMU remain on the tracks, mostly intact. 91 people including both drivers die in the collision, 54 suffer severe injuries. Two of the survivors succumb to their injuries weeks after the accident, bringing the death toll to 93. The deafening crash that echoed across the nearby town marked what became the country’s worst rail accident ever.
Aftermath
Residents and responders from nearby Harmelen were the first people on site, soon joined by responders from Utrecht as the scale of the tragedy became clear. Nobody in the leading car of the EMU had had any chance to survive the collision, the train’s rudimentary crash-protection structures were hopelessly overwhelmed by the heavy express train’s high speed impact, reducing the car to an unrecognizable field of debris. Most survivors were found in the rear sections of both trains, which had remained on track and largely intact.
The investigation found that the signaling-system had functioned correctly and that the signal box crew had followed all rules and guidelines when setting up the path for the EMU. They managed to recover the speed-logger from the remains of the express train’s locomotive, which showed that the driver had passed the yellow signal at full speed, not reacting until he reached the red signal where he triggered an emergency stop. At that point, however, there wasn’t enough space left ahead of the points the EMU was using to cross the express train’s tracks to come to a halt. The system was set up under the assumption that train drivers would start to slow down at the yellow signal, which would bring down speeds enough to safely stop ahead of the crossover. Not doing that in a train as fast as the express train left no chance to stop in time once the train even only got near the red signal.
Due to the degree of destruction and the death of both drivers the investigation chose to recreate the express train’s behavior, assembling a near-identical train for driving- and brake-testing on the 14th of January 1962. The experiment showed that the express train’s driver had triggered an emergency stop approximately 95m/312ft ahead of the red signal, which fell in line with local visibility-conditions at the time of the accident. It also meant that the express train was travelling at 125kph/77.7mph where it was meant to be at approximately 30kph/18.6mph.
With both drivers dying on impact there was no way to know for sure why the express train’s driver ignored the yellow signal. The official explanation is that the thick fog made the signal hard to see, along with the driver perhaps assuming he had the right of way as his train usually did. The second part of that theory is speculative though. The schedule meant for the express train to pass the crossing ahead of the EMU, this was changed spontaneously when the EMU reached the crossover well ahead of the delayed express train.
After the accident different options to improve safety at such crossover-sections were considered. One was to have two “expect stop”-signals ahead of a red signal, meaning trains slow down either earlier or the drivers have a second chance if they miss a yellow signal. Another option was the introduction of dual-red, placing two red signals in a row if trains were meant to stop. Both ideas were rejected as they would significantly reduce the capacity of the rail lines. It was furthermore noted that train drivers might be more likely to disregard yellow or even red signals if they thought that there would be a second “serious” signal behind the one they passed. This, obviously, could have catastrophic consequences if drivers assumed a double-signal where there was only one.
A removal of crossovers like the one at Harmelen was also not an option, being too expensive and in some places not possible due to space-constraints. As such the only realistic way to reduce the risk of train-collision after disregarded signals was to integrate an automatic train control system into the signaling system, which could interfere if drivers failed to obey the trackside signals. At the time such a system was already undergoing testing on a handful of rail lines in the Netherlands, but neither the trains involved in the accident nor the crossover at Harmelen were equipped with it. The system, called ATB at the time and since the 1990s ATB-EG (“Automatic Train Control — First Generation”).
The system allows continuous observation of a train’s speed by feeding a 75Hz current into the tracks which is constantly turned on and off, creating one of seven different pulse-frequencies. Each of these frequencies refers to a certain speed limit, ranging from 40kph/25mph to 140kph/87mph. The steel wheels and axles of the trains close the circuit between the rails, creating a magnetic field detected by components aboard the train. A bell aboard the train sounds each time a change in frequency (and thus, the speed limit) is detected. If the train is moving faster than the current limit a different bell rings in the driver’s cab for up to 3 seconds, giving the driver time to slow down before the system triggers an emergency stop regardless of the driver’s input. This emergency stop can only be terminated once the train is stationary. While ATB cannot trigger an emergency stop upon passing a red signal it can keep trains from approaching red signals at more than 40kph/25mph.
After the accident at Harmelen the introduction of ATB was sped up considerably, leading to 100km/62mi of rail line having the system operational by 1970 and all main lines being equipped by 1992. Further gaps in the coverage were closed between 2001 and 2002. Despite vastly improving operational safety the system had a few flaws, such as only being operational above 40kph/25mph, not functioning above 140kph/87mph and only having 5 different speed limits in between, leaving gaps of as much as 50kph/31mph between some of them. The system was also vulnerable to malfunctions caused by interference from certain electric locomotives as well as magnetic track brakes, leading to the latter’s use being banned in the Netherlands. In the 1990s the second generation of ATB (called ATB-NG) was introduced, addressing most of those flaws. As of 2023 ATB-EG, itself capable of preventing accidents like the one at Harmelen, is in the process of being replaced by superior, more capable train control systems.
The introduction of ATB-EG made a repeat of the accident at Harmelen already just about impossible, as the train would have been slow enough to stop safely upon reaching the red signal. A repetition became completely impossible in the 1990s when the crossover was replaced by a so-called flyover, taking trains leaving or joining the Utrecht-Rotterdam rail line onto a bridge over the westbound track rather than having to enter it.
NS 1131, the leading two cars of its train and NS 700 (the leading four-car EMU) were broken up for scrap at the site, it’s unknown if the rest of the express train was repaired or scrapped elsewhere. The Class 1100 was highly unpopular with train drivers, being loud, drafty and quite crammed. After another unit of the type suffered a train collision in 1978 drivers protested for better safety, leading the NS to add “crash noses” ahead of the cab intended to eat up impact-forces. The last units of the type were retired by 1999, with only five units being preserved by museums.
The Mat ’24 passenger cars were retired by 1972, followed by the Plan-E cars in 1988, 8 years after the NS had introduced newer, safer ICR-type cars for express trains. This made the Plan-E cars the last type of rail vehicle involved in the accident to be retired, as the last Mat ’46-provided service had run in 1983. Today a sole two-car unit (numbered 273) survives as an operational museum-piece.
An official memorial for the victims of the accident was unveiled on the 8th of January 2012, the 50th anniversary of the tragedy, within sight of the site. The memorial consists of two black stone plates carrying the names of all victims along with their date of birth in gold lettering, flanking a small stone pillar which carries a plaque reminding visitors of the accident as well as a small stone torso symbolizing the victims. Looking through the gap between the black plates has visitors look right at the site of the accident. The plates listing the names were replaced a few years after their unveiling to fix three typos that had been discovered in the meantime.
The speedometer of NS 1131 also survived the decades, having ended up in private hands before being donated to the Dutch Railway Museum in 2021.
_______________________________________________________________
A kind reader has started 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.
