Background
Voorschoten is a town of 25618 people (as of 2024) in the western Netherlands, located in the Zuid-Holland Province 8.5km/5mi northeast of Den Haag and 42km/26mi southwest of Amsterdam (both measurements in linear distance).
Voorschoten lies on the “Oude Lijn” (“Old Line”, also called the Amsterdam-Haarlem-Rotterdam railway), a double to quad-tracked main line connecting Rotterdam with the Dutch capital. The line opened in sections between 1839 and 1847, making it the oldest rail line in the Netherlands. The Weesp-Leiden railway (opened in 1978) and the Amsterdam-Schipol railway (1986) provide a shorter connection, but the old line remains an important rail corridor for both freight and passenger services. The 86km/54mi line is completely electrified and sees operational speeds of up to 140kph/87mph. The section of the line where the accident occurred counts four tracks, numbered 1 through 4 from west to east.
The Vehicles Involved
Train number 47612 was a mixed freight train operated by DB Cargo from Sittard via Rotterdam to Beverwijk. The train consisted of 26 freight cars in total, split into five type Shimms tarpaulin cars loaded with steel coils, twelve unloaded type Sahmms flatbed cars intended for steel coils and nine type Talns enclosed ballast cars carrying quicklime. The train was pulled by DB 189 054, a Siemens ES64F4 four-axle multi-system electric freight locomotive. The Class 189 was introduced in 2002, measuring 19.58m/64.2ft in length at a weight of 87 metric tons. Being a multi-system locomotive means it can operate in countries with different electrification standards, no longer requiring the locomotive to be swapped out at the border. The type has a top speed of 140kph/87mph, which is plenty for most freight services. As usual in Europe the train only carried a driver.
IC 1409 was an overnight Intercity passenger service provided by the NS (Nederlandse Spoorwegen, Dutch national railway) from Utrecht main station to Rotterdam main station via Amsterdam and Leiden main station. The service was provided on the day of the accident by a NS VRIM-set, unit number 9405. The VIRM (“Verlengd InterRegio Materieel”, “Lengthened interregional rolling stock”) is a four-car bilevel electric multiple unit (EMU) introduced in 2002, both by expanding the three-car DD-IRM-type trains and building new four- and six-car versions (the unit involved in the accident had four cars). Unit 9405 was part of the first batch of converted trains and once finished measured 107.56m/353ft in length at a weight of 235 metric tons. The trains can carry 407 passengers in a two-class configuration, including 29 folding seats, at speeds of up to 160kph/99mph. The train was almost empty at the time of the accident, carrying just 50 passengers along with a driver and a two conductors.
A construction crew working around Voorschoten station on the day of the accident utilized several bimodal vehicles for their tasks, including an excavator which was used as a compact crane to carry heavy materials. Bimodal vehicles exist in a variety of forms from cars to construction vehicles to firetrucks, with the bimodal name referring to a set of train wheels that can be lowered onto the rails. This allows usage of the vehicle along rail lines with far more safety than a rubber-wheeled counterpart, although the systems do tend to increase the weight notably. The excavator used at Voorschoten on the night of the accident was a Liebherr A 924 Rail Litronic, a 23 metric ton wheeled excavator measuring 9m/29.5ft in length. The excavator was driven by a 65 years old operator on the night of the accident, with the second seat (intended for an employee with a train driving license) being empty as the operator was sufficiently licensed.
The Accident
The Royal BAM Group, a Dutch construction company, was carrying out rail replacement work in several spots across Voorschoten station on the the night between the second and third of April 2023. Tracks 3 and 4 had been closed to traffic for the work, allowing trains to pass the site on tracks 1 and 2. Among the vehicles used at the work site was a bimodal excavator used to carry tools and materials across the work site from where they were stored to where they were required. The excavator was no longer required at approximately 3:10am on the fourth of April, so it left track 4 at a temporary level crossing south of the station. It was meant to be parked north of the rail line, so the operator radioed the local dispatcher for permission before, by 3:21am, he started making his way across the rail line.
Freight train 47612 is approaching Voorschoten station at the same time, travelling at 97kph/60mph in an attempt to catch up a 14 minute delay. The driver of the freight train spots the excavator in the distance as he nears the work site before realizing that the excavator is standing sideways on track 2. He triggers an emergency stop and lays on the horn, but there is no way his 1400-ton train is coming to a stop in time. The locomotive of the freight train crashes into the rear left of the excavator at 3:23am, still going at 91kph/56.5mph. The severely damaged excavator is thrown across the platform of Voorschoten station, ending up on track 1 along with rubble from the damaged concrete platform between tracks 1 and 2. The damaged locomotive comes to a stop another 360m/397yd down the line, just past the station. The shock of the collision incapacitates the driver for a moment before he tries to report the collision, only to find that his radio is out of order. He sets the locomotive lights to “emergency”, which should turn the lower two headlights red and the upper headlight white, but the lower left headlight was destroyed in the collision and a piece of the excavator is obscuring the upper headlight. The effect of the signal, intended to warn oncoming trains, is thus almost nonexistent.
The freight train’s driver uses his cellphone to call DB Cargo’s transport coordinator in order to report the accident at 3:25am, just as the headlights of an oncoming train come out of the night. It’s IC 1409 on its way to Den Haag. The driver of the Intercity knew about the construction site, but he was cleared to pass on track 1 at 140kph/87mph. He speeds past the locomotive of the freight train when he suddenly spots a yellow object in his path, a second before he crashes into the remains of the excavator at 133kph/83mph. The collision with the excavator and the platform’s debris derails the leading car of the train to the right, crashing through a gate as it turns sideways, mowing down everything in its path and and ramming a parked sprinter van (belonging to the construction crew) into the roadside ditch. Its speed carries the leading car across the ditch as it rolls over, separating from the rest of the train before coming to a rest upright in the adjacent field. The second car turns sideways too, overtaking the leading car and making it almost as far into the field. The rear two cars remain somewhat aligned with the rail line, derailing as well. A fire starts in the wreckage of the thankfully unoccupied van, which is eventually extinguished by construction workers. The excavator’s operator is killed as his excavator is obliterated by the freight train, while the driver of the passenger train and fourteen of his passengers suffer severe injuries. Another 15 people, including the freight train’s driver, are listed as suffering minor injuries.
Aftermath
The driver of the passenger train attempts to raise the alarm despite his injuries, only to find that the accident rendered his train’s radio inoperable as well. Unbeknownst to him local residents already called the emergency services at 3:25am, reporting a loud crash and, in some cases, yet another one as they were on the phone when the passenger train ran into the wreckage. The calls get more precise by 3:27am, when passengers aboard the train start calling about their train hitting something and derailing. A police car arrives at the site at 3:31am, followed by the fire department 6 minutes later and finally an ambulance by 3:44am. The alarm was continuously escalated, until a GRIP 3 was declared, indicating a risk to the well-being of over 50 people in one location. Simultaneously, the first helicopter reached the site at 3:53am. Most passengers of the Intercity were assembled by responders and local resident until the NS could organize busses to take them to Den Haag while, in total, 19 people were taken to hospitals by ambulance or helicopter.
The initial suspicion fell on the excavator being erroneously being used on the wrong track, theorizing that the construction crew had accidentally placed the excavator on a track they assumed to be closed. This has happened before in the past, usually in the form of construction crews confusing the two tracks of a double-tracked rail line for each other. Bimodal vehicles aren’t necessarily picked up by the signaling system, hiding such errors until a train driver spots the vehicle in his path. At that point, especially at night, its often too late to stop the train in time. However, this theory was soundly rejected by witnesses among the construction crew who managed to provide a decently detailed timeline of the excavator’s movements which was backed up by the radio traffic placing the excavator south of the rail line minutes before the accident. In fact, the excavator hadn’t even been used on track 3, the one closest to the active tracks, that night.
Protocols and witness accounts from the work site showed that the excavator had been put into service at the site at approximately 1:50am. This had required a brief closure of tracks 1 and 2, as the excavator had been parked north of the rail line since the afternoon. The construction crew’s head of operational safety radioed the dispatcher at 1:31am and asked for tracks 1 and 2 at the site being closed, being told to wait approximately 10 minutes. The dispatcher gave permission to “disable” the two northern tracks at 1:43am, which was done by placing a pair of “Short circuit lances” between the rails. These folding metal tools connect the rails of a track similar to how a train’s axle and wheels would, creating a connection for the low current sent through railway tracks. This closed circuit is picked up by the signaling system, registering the section as “occupied” which prevents trains from entering that section of track. They serve as a secondary safety measure, assuring that no train can approach a closed track even if the dispatcher would forget about the closure. These lances were used somewhat permanently on tracks 3 and 4, closing them on a large section across several work sites, and were also used to temporarily close a smaller section at Voorschoten station to have the excavator and a “welding bus” (a bimodal sprinter van carrying welding equipment) access the work site.
The investigation managed to recover a data logger from the remains of the excavator, showing a switch to “rail mode” at 1:48am. This indicates that the excavator had reached track 4 at that time, supporting the protocol and claims that tracks 1 and 2 were reopened at exactly that time by removing the lances and reporting as much to the dispatcher. The excavator was used at various work sites along track 4 over the next hour, before being sent back to the temporary crossing south of Voorschoten station by 3:10am. The operator drove it the 3.5km/2.2mi back to the crossing alone, greeting coworkers who were working on track 3 as he passed them. This encounter again confirmed that the excavator was on the correct track as it reached the crossing.
The excavator reaches the crossing at 3:21am, switching from rail mode back to road-mode. This lifts the rail wheels off the ground and switches the hydraulic powertrain back to driving the regular wheels. The head of operational safety catches up with the excavator at the same time and contacts the dispatcher, asking for permission to close tracks 1 and 2 to move the excavator and a welding bus across the tracks to their storage position. The dispatcher tells the construction crew to wait for ten minutes before the head of operational safety contacts his coworker at another work site a short distance down the tracks to discuss proceedings there. This presumably took his attention away from the excavator. It is not 100% certain what happened between the instruction to wait for clearance and the fatal collision just 2 minutes later, since the two men were alone at that section of the site and the operator of the excavator didn’t survive the accident. The investigation thus had to largely rely on the witness account of the one surviving person at the site along with GPS- and motion data from the excavator’s data logger and an examination of the wreckage. The excavator’s data logger records the excavator starting motion in road-mode at 3:22am. The excavator then stopped almost immediately, followed by a sudden rotation being recorded along with an instant acceleration to 3kph/1.9mph at which point the recording shuts off. It’s assumed that the rotation and acceleration stem from the impact by the freight train, which immediately cut power to and/or damaged the data logger (which is why the acceleration is only documented up to 3kph).
The investigation consulted Liebherr, the manufacturer of the excavator, to assist in the examination of the wreckage. No sign of a defect that could have led to the accident could be found, and the service- and maintenance history of the excavator were without cause for concern too. Similarly, the coroner couldn’t find any sign of a medical condition that would have incapacitated the operator ahead of the accident. A theory that he inadvertently set the excavator in motion was thus also rejected, as was one that implied the excavator did so on its own. The police looked into the background and life of the operator, but ended up ruling out a suicide as it was deemed extremely unlikely. A malicious act was also ruled out for lack of evidence. Comments from coworkers about high workloads were also considered, but since it was the end of the shift there was no sign that the operator had to quickly park the excavator and move on to the next task. Coworkers testified that the operator had seemed relaxed and calm ahead of the accident, being in no rush as he returned to the crossing. The recorded motions in between stopping on track 4 at the crossing and the data logger failing as the initial collision occured also happened in a pace that didn’t indicate a rush. Coworkers further pointed out that they had all been aware of trains passing the site at regular speeds, explaining that none of them would attempt to cross the active tracks as they knew how little time there was between spotting headlights and the train passing.
All this leaves little explanation beyond momentary human error in the form of temporary confusion, false interpretation or miscommunication. The report explains that the investigation settled on one of these factors being the primary cause, but that an exact definition and explanation cannot be reached. It’s possible that the driver misunderstood the instruction given to him regarding the wait time ahead of crossing the tracks and thus crossed them early, but it’s just as possible that he got an instruction that wasn’t ideally worded which led to him misinterpreting it, or that he, for unknown reasons, assumed that he had been cleared to cross without having actually received clearance. The evidence and indications are insufficient to declare one of these possibilities as the definitive reason for his actions, leaving it on “human error”.
The report also dedicates two lengthy chapters to contributing factors, which played a rather big role in the accident. The evaluation of these factors was based on an examination of 1200 prior situations where there had almost been accidents, 120 highly similar situations among which were closely examined. First and foremost is the infrastructure at the site, which put a two-track work site on the southern half of a quad-tracked rail line while only being accessible from the north. This meant that any access to the work site required crossing two railway tracks. This usually wouldn’t be much of a problem, but in the case of Voorschoten the decision was made to leave the two northern tracks fully operational aside from very brief temporary closures to transfer equipment, minimizing delays and cancellations for the price of having trains pass between the access road and the work site at rather high speeds. This setup actually presented an improved situation which was realized after past accidents.
It had been standard practice in the Netherlands until the 1990s to conduct maintenance work on active tracks themselves, with the workers briefly moving off the tracks whenever a train passed. The report notes that this practice led to three fatal accidents in 1995 alone, which caused a change in the risk assessment of maintenance- and repair work on rail lines. The argument of merely lowering the speed of trains passing through in favor of not pressuring workers into quickly finishing their tasks on a fully closed line was seen as no longer sufficient and a reliable separation of rail traffic and maintenance crews had to be implemented. The result was a decision that single-track lines would now be completely closed to traffic while workers were active on the line, while double-tracked lines would see the track being worked on shut down. Dual safety measures also had to be implemented to keep trains from entering the work site unintended. This was the reason for the lances at Voorschoten, being a secondary safety measure beyond the official closure since they ensured that the track would be registered as occupied, keeping signals red and prohibiting a train from being sent into the work site.
The rule was upgraded to double-tracked lines being completely shut down for work on one of their tracks in 2012, after several near-miss incidents and finally a fatal train collision had shown that the existing tactic was insufficient. The fatal accident had occurred on the 21st of April 2012 at Amsterdam-Westpark station when a train entered a section where oncoming services had to take turns using a single track without authorization, colliding head-on with an oncoming train. The trains had had to take turns because the line’s second track was closed for maintenance work, and the train that ran the red signal didn’t usually encounter a red signal there, surprising the driver. The accident injured 116 people and claimed the life of a passenger.
The upgraded safety measure explicitly addressed double-tracked sections, excluding lines with more tracks as it was figured that those would always leave one dedicated track per direction available. Maintenance work is furthermore usually performed on one track at the time, which would leave a one-track “buffer zone” on quad-tracked sections to keep tools and workers out of harm’s way. The report on the accident at Voorschoten explains that, while this still creates a risk where the trains switch over to the two open tracks along with a risk for workers and passengers if a work site occupies more than one track those risks are subject to individual risk assessments and don’t justify standardizing a more complete closure of lines with at least four tracks. According to the official explanation closing a four-track rail line for work on one or two tracks creates an unreasonable negative impact on train operations compared to the increase in safety. In fact, a risk analysis by the NS in 2022 came to the conclusion that directing traffic past a work site on a single track for both direction shouldn’t be completely banned but could be allowed in selected cases, largely undoing the upgrade after the Westpark collision 10 years prior. Furthermore, the accident at Voorschoten occurred while reintroducing limited traffic to tracks that are being worked on was being evaluated.
The report summarizes that the safety improvements of the recent history didn’t sufficiently consider both multi-track work sites and work sites on quad-tracked rail lines (such as the one at Voorschoten), which, while keeping trains on one track per direction, create the same risk for workers by the trains and for the trains by the work site as passing a work site on the other track of a double-tracked rail line. The fact that the work at Voorschoten took place at night further elevated the risk involved, but this, too, apparently wasn’t sufficiently considered. Neither was the option of letting trains pass at reduced speed. In general, it appears that the risk-assessment ahead of working on rail lines also didn’t properly consider the risk to trains by the work site, focusing on the risk between trains and posed by trains for the workers. Trains have been derailed, sometimes with fatal consequences, by much smaller obstacles than the excavator, and it has to be considered lucky that the freight train remained on its rails. The thought of it derailing, possibly into the path of the oncoming passenger train, bears horrific possibilities.
Beyond that, while there was sufficient effort taken at Voorschoten to keep trains from entering the work site, there was no proper way to keep workers or their equipment from ending up on the active tracks. Workers were told that two tracks were closed and the other two were fully operational, but there was no physical barrier or even indication of the edge of the work site. Nothing prevented a worker or their equipment from ending up in the path of a train, there wasn’t even something like barrier tape to warn workers of the active tracks next to them. This, especially in the nighttime conditions, posed an unnecessary risk. The report also points out how the operator of the excavator and the dispatcher were never talking directly, always going across at least one more person which increased the risk of miscommunication. This is another part that is escalated by the nighttime setting, increasing the risk of exhaustion and fatigue as contributing factors through lowered focus. Anyone who’s played the Telephone game as a child knows how easily a message can change when going through several people even if they aren’t exhausted or fatigued.
The report closes with several recommendations for several recipients. The infrastructure ministry is advised to ensure that ProRail, the entity in charge of Dutch rail infrastructure, places higher value on work site safety and safe operations around construction efforts, moderating/ending the unreasonably prioritization of keeping trains running and being more open to accepting a higher negative impact on rail service in return for safer maintenance and construction operations. ProRail themselves were recommended to create a system to record and archive near-accidents in order to make information about them easy to research, and to use the gathered data to improve operational safety regarding inspection and maintenance efforts.
They were also recommended to specifically improve the safety of all rail workers surrounding maintenance- and construction work sites, especially if not all tracks could be closed down. They were strongly recommended to:
- Ensure the existence of a robust physical barrier between active and closed tracks.
- End the practice of “island closures” like the one at Voorschoten, meaning that workers will no longer have to cross active tracks to access the closed tracks/work site.
- Ensure the existence of a visual indicator to clearly tell workers if a track is closed or in operation.
- Have any safety-relevant conversations at work sites recorded for future reference, not just those with dispatchers.
Lastly the investigation demanded that the negative effects of nighttime work and overtime on workplace safety and workers’ health are given more attention when planning future projects, with the goal of largely eliminating overtime for nighttime work on rail lines and ensuring all present individuals have a sufficient rest period between shifts. This part of the recommendations came to be after several workers, not just from the site at Voorschoten, told the investigation that their work- and rest-hours were “mostly theoretical”.
The official report was released in May 2024, listing “human error” by the deceased excavator operator as the main cause. The main lesson to be learned, according to the report, was that maintaining the intensively used Dutch rail network can and has to be done in a safer way than presently conducted. No other person involved in the accident was found to be of any individual guilt, leading to the criminal investigation ending with no charges being filed.
The freight train’s locomotive was towed back to Germany in June 2023, while the passenger train left the site on flatbed trucks. It’s currently unknown if the locomotive or any of the passenger train’s cars will be repaired.
The line reopened on the 20th of April 2023, after the tracks and catenaries were replaced and Voorschoten station was repaired. The site looks like it did before the accident, aside from a few new tiles in the station platform. A trio of trees was planted near the site of the accident, surrounding a a steel plate showing a black and white photo of the passenger train’s wreckage and the date of the accident. In what might be a first for a memorial a QR-code on the image leads to a page by the municipality explaining what happened at the site, rather than having an inscription on the memorial.
_______________________________________________________________
The original report, which goes into more detail on the contributing factors, can be found right here. It’s 178 pages and exclusively in the Dutch language.
_______________________________________________________________
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.
