Landing Crashed: The 2023 Stockholm (Sweden) “Jetline” Roller Coaster Derailment
Back when my blog reached the 200th post I “diverted” from the usual topic and covered the 2015 train collision on the “The Smiler” roller coaster. The article was quite well received, so I decided to do it again. The blog isn’t at 300 yet, but 250 seems like a decent milestone too. It’s still a train running on (and derailing from) steel tracks, just not a railway-train. And to those wondering, or those who don’t care about this kind of train, the blog will be back to the usual kind of trains next month. So, here’s post #250:
Background
Stockholm is capital of Sweden, home to 995574 people (as of 2025), and is located in in southeast Sweden 400km/248mi northeast of Göteborg and 265km/165mi southwest of the Finish city of Turku (both distances in linear distance).
The city is home to the Gröna Lund (Swedish for “Green Grove”) amusement park established in 1883. The park offers various theaters and stages for events along with restaurants and various amusement park rides, crammed into a comparatively small footprint filling the peninsular of Djurgården. One of the park’s highlights for decades was “Jetline”, a steel roller coaster which opened in 1988 and was refurbished/modified in 2000. The coaster offered a track length of 800m/2625ft with the highest point reaching 32m/105ft after the refurbishment. A lap took 90 seconds, allowing a theoretical 1300 passengers per hour to experience speeds up to 90kph/56mph and 4.5g (4x the earth’s gravity, meaning a 50kg person felt a weight of 225Kg pulling on them for brief moments) aboard one of the three trains. The coaster featured a distinct, excessive support structure, which was constructed to carry a scenic mountainside-facade which ended up never being constructed. Two sections with brakes limit the trains’ speed partway through the lap, another slows trains down right ahead of the station.
Each train consisted of seven cars, each housing two people, and was made of a steel frame with glass fiber reinforced plastic covers. Each train’s first car had a weight of 786Kg, while the other cars (being slightly smaller) weighed 500Kg. As such a whole, empty train hada weight of approximately 3800Kg, being calculated to reach up to 4900Kg if it was loaded with 14 people. Passengers were secured into the padded bench with a steel lap bar, which was lowered into their lap before the train left the station and was meant to ensure the passenger could not get up or unintentionally slip out of the car due to the G-forces. Lap bars are usually preferred by enthusiasts to restraint-systems going over the shoulder, as they are less restrictive.
As with most coasters the trains were unpowered, getting their speed from a chain dragging them from the station up an incline to the coaster’s highest point, where the train is released and completes the circuit with momentum alone. The train rides on several six-wheel bogies, with a pair of wheels on each bogie running on top of the tubular rail, a pair sitting on the outside (limiting lateral motion) and one pair sitting under the rail (called “Upstop” wheels) to keep negative G-forces (such as at a crest) from lifting the train off the tracks. Each wheel is made of steel with a solid plastic tire, and the forward-most pair of bogies was attached to a wide wishbone attached to the underside of the car with a hinge allowing some movement around the longitudinal axis (the other bogies were attached rigidly to the train).
The Accident
The 25th of June 2023 was a nice summer day, with clear skies low wind and a temperature of 26° Celsius/79°F before noon. A train departed the station at 11:36am with 11 passengers onboard, leaving one empty spot each in the first, fourth and sixth car. The staff at the station would later say that everything seemed fine as the train arrived at the station, was loaded with passengers and then departed up the first hill, but some passengers who’d ridden the coaster earlier in the day recalled that something had felt “off”, that the ride felt bumpier and created more noise than usual. It wasn’t really suspicious though, because Jetline was an old all-steel coaster, so of course there’s some noise when it runs.
The chain pulled the train up to the highest point, releasing it into a downhill right hand turn followed by a straight downhill section into an artificial tunnel, housing a wide 180° right hand turn. The train climbed steeply uphill again, navigating a tighter 90° right hand turn into a brake-section at the second-highest point of the layout. By the time it departed the brake-section into a tight, steep downhill right hand turn the train was doomed. The 180° turn brought the train nearly down to ground level, at which point the leading car’s suspension failed, allowing the chassis to scrape along the track. The train still carried enough momentum to start the following right-hand ascend, at which point half of the leading car’s wishbone departed the car along with the right-hand bogie.
The train still went over the crest of the following hill, now rattling along the tracks. The drop after that hill brought the first harder strike against the track, shaking the whole coaster. This impact didn’t stop the train, but was enough to eject a passenger from the fifth car. The remaining part of the wishbone soon separated from the leading car along with the left hand bogie, leaving the train to essentially drag the front end of its leading car along the tracks. Two passengers from car seven subsequently fell from their seats, with one of them falling to the ground while the other managed to grab a piece of the support structure and hold on. The slowed train proceeded through the valley ahead of the next hill and started the ascend, but ran out of momentum and rolled back into the valley, finally coming to a stop. The passenger who had been ejected first died when they hit the ground, while 3 passengers (the second person to be ejected and 2 who remained aboard the train) suffered serious injuries. The passenger who had been ejected but grabbed the support structure was retrieved by firefighters with minor injuries (as were 6 passengers who remained on the train).
Aftermath
Staff at the coaster had initially notified maintenance personell when the first piece of the wishbone had fallen from the train, calls to first responders were placed by both staff and visitors at 11:38am with firefighters, ambulances, emergency doctors and police units being dispatched to the park. The first responders to arrive where firefighters who reached the coaster by 11:45am, starting by rendering first aid to the passengers who had fallen to the ground and ordering the ladder truck’s crew to begin accessing the survivors trapped on the train or on the support structure. An evacuation of the park was also initiated a few minutes after the firefighters arrived.
A technician from the park secured the leading car to the track with ratchet straps and managed to release the lap bars keeping passengers in their seats, drastically simplifying their rescue as the firefighters now didn’t have to cut each bar individually. The trapped passengers were provided with water due to the hot weather and the last passenger finally reached the ground via the ladder truck’s basket just over an hour after firefighters arrived. The other two trains of the coaster had been at ground level, allowing maintenance personell to evacuate those without help from the responders.
The direct cause of the accident was quickly determined, the entire front suspension of the train had gradually departed the car, causing the train’s frame to repeatedly impact the tracks. The coaster track consisted of a large tubular “spine” connected to the two tubular rails by rectangular braces at regular intervals. Impacting these braces along with the joints between track-sections had caused the repeated impacts, similar to a railway train being dragged along the sleepers between its rails.
Investigators examined the track of the coaster with a boom truck and a drone, finding light scratches in various places but also more severe impact damage to the track and its bolts in selected places. A crane was brought in to remove the stranded train from the track once it had been examined on-site. The removal required cutting the train apart in some places as some of the couplers had been damaged beyond operation during the derailment. Impacts between the cars, permitted by the damaged couplers, had also bent the forward wall of several cars backwards towards its passengers.
When car 1 was transported to an off-site facility to be examined the investigators found severe impact-damage to its stell frame near where the bogie, of which only a small piece had remained on the train car, had been located. This damage matched impact-marks on the track.
The focus of the investigation soon went to the wishbone, with an engineering firm being tasked with the examination to try and explain the failure. In total the engineering firm was provided with half a dozen wishbones:
- The failed wishbone from the accident train, which had done around 5000 laps of the coaster
- An identical wishbone which had been installed on a different train and had done around 29000 laps
- Three new wishbones which hadn’t yet been used
- An original wishbone from 1986, to compare to the other 5 wishbones which had been manufactured as spare parts in 2019.
Both the old and new wishbones had been manufacturered to the same specification, making them effectively identical. They consisted of a trapez-shaped piece in the center (which, for the forward-most wishbone, held a lengthwise tube for the hinge-mechanism) welded two two rectangular pieces of hollow steel tube, each of which had a vertical round steel tube welded to the far end of it. The whole construction had a wall-thickness of 5.6mm and the rectangular sections were 100x100mm/3.93x3.93in. On the train involved in the accident the wishbone had come apart at the welds, breaking down into the components it was constructed from.
The engineering firm started by examining a slice cut from the the center part of the failed wishbone which had remained attached to the train throughout the derailment, examining the fracture-surface to either side of the trapez-shaped section with a stereo-microscope. The examination showed that the weld only involved about half of the wall’s thickness and that there were “gaps” in the added material’s adhesion to the surface, meaning the weld wasn’t actually going all the way around the circumference as it was meant to. In isolated spots the engineers even still found groves from when the part’s surface was sanded in preparation for the welding. Such defects obviously severely affect the rigidity of the assembled part, both in static strength and fatigue resistance.
The fracture-surfaces were also examined with an electron microscope, finding marks on the fracture-surfaces which indicated that a fatigue-crack had started from inside the wishbone and begun working its way through the already insufficient weld. The investigation proceeded by cutting identical samples from a wishbone manufactured in 2019 which hadn’t been used yet and one which had been manufactured in 1986. The cutting-process alone found a difference between the old and new wishbone, with a steel plate within the rectangular section acting as additional bracing in the original wishbone while no such plate was found in the 2019-wishbone.
Cross-sections were subsequently cut from welds on both the original wishbone and one of the newer ones, each wishbone “donating” a cross-section from a corner of the rectangular tubing and from a flat side. The welding in the corner was faulty on both wishbone-samples, but the ones on a flat side was perfectly fine in the old wishbone and showed insufficient fusion on the new wishbone, reducing how much of the surface-area was actually connected. The welding on the new wishbones was also so uneven that no properly flat surface for adhesion had been created, further reducing fatigue-resistance. A fatigue-crack was even found on the corner-weld, spreading from the inside of the tube and starting to increasingly weaken the wishbone. Shockingly, that wishbone had been in use since 2022.
The examination of the samples further showed lower quality finish on the surfaces before welding, with the new wishbone’s surfaces being at a right angle instead of being beveled. The diameter of the walls right at the weld-surface was also thinner than the rest of the tube in some spots, due to uneven surfaces and sharp edges. This meant that the part where the weld attached to the tube (poorly so, too) was the weakest part of the tube. The absence of the reinforcement plate in the tube yet further reduced strength of the piece and made executing a good weld more difficult as the material was more likely to flex/shift under the heat introduced by the welding process. The absence of the plate in the new wishbone was not explained in the report, it’s likely that the investigation didn’t receive an explanation. The engineering firm did note that the reinforcement was absent from schematics provided to the company which manufactured the wishbone by the park. This contradicted a statement from the park claiming the reinforcement was in the provided schematics.
The next part of the train to be examined were the lap bars. Their purpose was to keep the passengers securely in the cars no matter what happens, which they obviously and tragically failed to do during the accident. In normal operation the bars are “open” (moved to a near-vertical position) when passengers board the train and are then pulled into the lap of the passengers, using a ratchet-system to lock them in the desired position. They each consist of a steel tube in the shape of an angular inverted U, with foam padding on the transverse bar which comes in contact with the passenger. Once the ratchet-system engages the bar cannot be opened by the passenger, the system is unlocked either by a “Hook” at the beginning of the station hitting a release-button on the outside of the car or by a special hand-tool available to the staff for emergencies (such as the evacuation after the accident). Once unlocked a spring-mechanism makes the bar swing up into the near-vertical open position.
During the accident several lap bars were bent out of shape, including those where people had been ejected being found in a near-vertical position. Neither these bars nor those in unused seats had been touched during the rescue and recovery operation, allowing a good indication of their position and function during the accident. The locking-mechanism was found to be intact on all cars, and all of them had been properly lowered and locked ahead of the accident.
The investigation proceeded to source an undamaged car from a different train and removed the front from it, allowing free access to the lap bar system. The car was then secured in place and the locked bar was pulled on with a chain-hoist as the force used was continuously measured. The test was performed with the chain pulling from different angles, with the ratchet-system set to either nine notches (similar to how it would be set for a child) and three notches (large adult). Both settings yielded the same result, with the bar starting to flex slightly but returning to its position once the chain relaxed before permanently deforming at 1400 newtons (just over 140Kg of load) similar to how they had been deformed in the accident. The permanent deformation occurred before the ratchet system would fail. A later test run would show an “adult” dummy exert no more than 31.3Kg of load on the bar during an emergency stop. This gave an indication of how far beyond the predicable loads the bars had been taken as passengers hit them during the derailment. It was concluded that there was no fault with the restraint-system, as their deformation only took place once exposed to loads well beyond what is to be expected in operation, even in emergencies.
Attention thus turned back to the wishbones, which had been manufactured in 2019. Their manufacturing was deemed necessary after fatigue cracks were found in the center of a train’s forward wishbone that year, affecting the welds holding the tube in place. General wear on the wishbone was deemed to be too severe to make a repair sensible and Gröna Lund chose to have a new set of forward wishbones made, covering all 5 trains. The park contacted the Swedish company Mekosmos AB for the new wishbones, a company that had manufactured other spares for the coaster in the past with satisfactory cooperation and results. Technical drawings were provided showing the detailed construction for the wishbones, with all measurements, angles and the exact way parts were meant to be welded together (“butt welding”, meaning the parts don’t overlap but are “end to end”). Mekosmos didn’t offer welding work themselves, so that part of the production process would be forwarded to GMW. Mekosmos did not tell the park that they would use a subcontractor, and with GMW’s quotation for the work matching what the park was willing to pay they ordered the new wishbones in December 2019. The worker who performed the welding-work told the investigation that there had been some uncertainty about parts of the production-process, but that they consulted with coworkers and agreed on how to do it without asking Mekosmos or the park. The worker held no official certificate (“welding licence”), but did have a lot of documented experience with welding different materials. GMW sent the new wishbones to an external radiology-firm to check them for defects and flaws and, once they returned with spotless results, forwarded them to Mekosmos for delivery to Gröna Lund.
Gröna Lund received the new wishbones in spring 2020, with no checks being performed upon reception as it was planned that they would be examined before being installed on the trains. A visual examination of an unused wishbone by the investigation showed damage near the tube of one of the wishbones, which had likely occurred during the manufacturing process as the damage was beneath the wishbone’s paint. When asked about the damage Gröna Lund claimed that, while the wishbone had been stored with any other new wishbones, the damage was known and the wishbone was meant to be discarded when it was picked up by the investigation. This claim could be neither proven nor disproven.
In summer 2022 workers noticed that a middle car on one of the coaster’s trains was out of alignment, leading to immediate retirement of the train from active use. An inspection showed a sizable crack in the rear wishbone of the first car, which had spread to several milimeters wide and affected 3 sides of the wishbone’s square tubing. The coaster was shut down and all trains were checked for cracking, with rear wishbones on 2 more trains showing defects. DEKRA, a German testing, inspection, and certification company, was contracted to assist with the checks and was also consulted about how to handle the discovered cracks. The cracks were then welded up by licensed Gröna Lund staff under supervision of DEKRA-employees, with post-repair testing showing no anomalies in the repaired wishbones. The park still chose to order replacement control arms from Mekosmos, who once again subcontracted the work to GMW without telling the park. However, GMW had no capacities for the contract and thus hired yet another subcontractor to manufacture the wishbones for Mekosmos for the park. The first 15 new wishbones arrived at the park in March 2023, but failed visual inspections and were found to have “annormal” welds didn’t meet the specified measurements. They were thus sent back to be scrapped. The replacements for these replacement wishbones took so long to manufacture that they wouldn’t see installation before the accident. The trains continued running with the repaired wishbones instead. The repairs weren’t documented by the park, but the investigation found that they were traceable through DEKRA’s records, listing 12 welding-repairs between 2006 and 2023.
One of the new forward wishbones had been installed on train E (the train which had had the accident) in March 2023, at which point the train had been out of service for several months after a crack was discovered in a rear wishbone in August 2022. The intention had been to replace the rear wishbone with a new one as well, but due to those arriving defective and requiring a return it was instead replaced with an unused spare wishbone from 1988. The DEKRA performed so-called Eddy Current Testing (ECT) on the train’s wishbones in spring 2023, a non-destructive electromagnetic test to find surface- and near-surface flaws in conductive materials (such as the steel wishbones). The testing was without any suspicious result. Train E thus reentered service in late April 2023, from which point on only the regular visual inspections were conducted. Nobody seems to have noticed at any point that the new forward wishbone differed significantly from the originals, most glaringly with the absence of an internal bracing plate. This, along with faulty welds, had significantly reduced the strength and rigidity of the wishbone compared to the originals, but in a way that wasn’t caught by the conducted tests.
The faulty wishbone thus wore and accumulated damage “from the inside out”, until, on that fateful June day in 2023, the cracking became so severe that the remaining material couldn’t withstand the loads exerted on it any longer and failed. The loss of (pretty soon) its entire front suspension assembly left the train unable to continue rolling along the track as designed. It started hitting the track between the rails at various forces depending on the location’s g-forces, with two impacts being hard enough to shake the entire structure. Two impacts with the track structure were so hard that the resulting forces overwhelmed the structural rigidity of the restraint system as passengers impacted them, bending them towards the “open” position. A passenger was fatally ejected when this happened, with two more passengers being unable to keep themselves within their car without the lap bar’s assistance on a following slower downhill section as the train had begun to lean severely to the right from the incurred damage.
The investigation was in no position to demand the park to do anything (such investigations are to find a cause, not decide fault and consequence), but the recommendations included asking Gröna Lund to improve their safety-systems so that future replacement parts are guaranteed to meet the required specifications, procedures are created to continually look for, identify and adress potentially dangerous risks in the organisational and technical operation and assure inspections of all attractions are tailored to the age, operating hours and risks of each attraction. Upon the release of the report Gröna Lund announced that Jetline would not reopen, seeking its removal. Lastly, they were fined 5.2 Million Swedish Crowns (480600€/559400USD) by a Stockholm District Court for fatal negligence, with GMW was fined 1.3 Million Swedish Crowns (120150€/140 thousand USD). Mekosmos was acquitted.
_______________________________________________________________
The rather extensive english-language investiation report can be found right here
_______________________________________________________________
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.
