Looped Out: The 2020 Wallan (Australia) Train Derailment

Max S
18 min readMar 10, 2024

Background

Wallan is a town of 15004 people (as of 2021) in southern Australia, located in the state of Victoria 104km/64mi east of Ballarat and 46km/29mi north of Melbourne (both measurements in linear distance).

The location of Wallan in southeast Australia.

Wallan lies on the North East railway line, a dual gauge partially electrified dual to eight-tracked main line connecting Melbourne with Albury (New South Wales). The line consists of tracks laid in two different gauges, making it available for both trains built in regular gauge (where the rails are 1435mm apart) and Australia’s broad gauge (where the rails are 1600mm apart). The line opened in sections between 1860 and 1891 as a purely broad-gauge line, almost 100 years before added regular-gauge tracks or reconfiguration of existing tracks enabled full-length travel by regular gauge trains.

Wallan is the site of the Wallan passing loop, a siding on the regular gauge track allowing both oncoming trains to pass each other and faster trains to overtake slower ones as there is only one regular gauge track at the site.

The site of the accident seen from above today, the train approached from the north (top of the image). The gap cut into the trees to recover the wreckage is still visible.

The Train Involved

Among the trains using the regular gauge side of the North East railway line is the New South Wales XPT (eXpress Passenger Train), a long distance express passenger service provided by NSW Train Link, a government-owned rail service provider focused in the southeast of the continent. Each XPT unit consists of four to seven four-axle passenger cars in between two diesel locomotives, running in fixed consists similar to other countries’ high speed trains (which is why the locomotives are usually referred to as “Power Cars”).

An overview of a 7-car XPT, created by New South Wales’ transport agency. The unit involved in the accident was a 7-car set as well.

The XPT was introduced in 1981, being based off the British InterCity 125 high speed trains (commonly called the HST) with some modifications to accommodate the conditions in Australia. The key differences between the British and Australian power cars are shorter overall length, different suspension, lower engine output and improved cooling while the passenger cars are an independent design making use of the bigger loading-gauge found in Australia to offer more interior space. The trains are capable of reaching 200kph/124mph, but are geared to top out at just 160kph/99mph instead.

The unit involved in the accident, led by XP2018, carried 155 passengers, five crew members and an “accompanying qualified worker” (AQW) at the time of the accident, being driven by 54 years old Mister Kennedy.

An 8-car XPT-unit, identical with the train involved (except for one extra passenger car) photographed in 2009.

The Accident

ST23, an XPT service from Sydney’s Central Station to Melbourne’s Southern Cross Station, departs Sydney at 7:41am on the 20th of February 2020, just a minute behind schedule. It’s meant travel across New South Wales and Victoria before ending its southwest-bound journey at Melbourne on Australia’s southern coast by 6:30pm. An uneventful journey sees the train reach the city of Albury (New South Wales) at 4:37pm. The passenger service crew is replaced at that station, with a new crew taking over for the section of the trip through Victoria. Mister Kennedy had taken over from a different driver at Junee, about 2 hours prior.

The train’s route as shown in the official report, I reinforced the marker for the accident-site to make it more visible.

ST23 departs Albury at 4:44pm, 1 hour and 29 minutes behind schedule. The crew would later describe the journey as perfectly normal, apart from the delay causing some frustration among the passengers. Mister Kennedy is contacted by the local traffic control center (run by the Australian Rail Track Corporation, ARTC) soon after departing Albury, notifying him of a nearby alert showing a train passing a red signal and informing him that ST23 would be sidelined at Kilmore East to let an oncoming regional train pass. The transmission also includes the traffic controller telling Mister Kennedy that the train would be sent through the passing loop at Wallan. Mister Kennedy confirms the information received, but makes no mention of going through the passing loop at Wallan.

The decision to send ST23 through the passing loop has its origin in an incident 2 weeks prior, when a road vehicle had gotten caught on the rail line’s overhead wires at Wallan, causing a power surge in the nearby equipment shed which started a fire, leading to extensive damage. The fire disabled both signals and level crossings in the area, including the signals indicating if a train was about to be directed into the passing loop or could stay on the main track. The issues forced ARTC to impose various safety measures including a 25kph/15.5mph speed limit, which obviously led to massive delays. A new set of measures was thus implemented in the evening of the 7th of February, which included disabling the points leading into Wallan’s passing loop so trains would always stay on the main line, allowing them to maintain speeds as high as 130kph/81mph. The disuse caused the loop’s rails oxidize, leading to the decision to temporarily send trains exclusively through the loop between 2:30pm and 9:30pm on the 20th of February. The signaling system was still down, in fact “cleaning” the rails by running trains over them was required to start post-repair testing of it. The points had a speed limit of just 15kph/9mph when entering the loop, little more than a tenth of what they could handle in the “straight ahead” position. Train drivers had been issued a notice about the loop being used for those seven hours a few hours prior.

A graphic from the report showing notable locations around the site of the accident. Kilometers listed are distances from Melbourne. Note that North is to the left.

ST23 arrives at Kilmore East at 6:56pm to let regional train 8625, which had been the second train to go through the passing loop, pass before it can enter the 24km/15mi section where the special safety measures are in place. At the same time a crew of rail workers is preparing for ST23’s arrival at the Kilmore East passing lane signal (KME16), including Mister Meintanis, the AQW who will ride along in the driver’s cab between the signal at Donnybrook. Mister Meintanis had been handed a printed copy of the Traffic Notice 367 (notifying drivers that trains would be sent through the loop that day) at the start of his shift, and ST23 would be the shift’s first train to be accompanied by him. His main job, however, was not to advise the driver of the passing loop being used that day, but to ensure that another worker manually activated a level crossing just south of the passing loop as they approached. ST23, having finally been passed by the oncoming regional train, reached KME16 at 7:31pm and was boarded by Mister Meintanis, who, in the meantime, had been given further paperwork by the local signaler operating KME16. He now carries the notice about using the passing loop and the associated speed limit (TN 367), the individual permission (“Train Authority”) for ST23 to go from KME16 to Donnybrook (TA 17), and a warning-note regarding the level crossing (CAN 7). The contents of the conversation between Mister Kennedy and Mister Meintanis aren’t known as the locomotive wasn’t fitted with a voice recorder, but a radio call from the traffic controller at 7:32pm sees Mister Kennedy reply “yeah, authority 17 and CAN number 7 filled out ahh the same way it has been for the … rest of the time“ when asked if he had received all the paperwork from Mister Meintanis. He did not mention TN 367, and there was no inquiry about it by the controller.

ST23 departs KME16 at 7:34pm, entering the single-track section to Donnybrook. The control center knows that the points are set for the loop, that Mister Meintanis aboard ST23 knows that, and they assume that Mister Kennedy knows too. Mister Meintanis calls the level crossing keeper (LCK) at 7:41pm, ordering the activation of the crossing and receiving confirmation that the crossing sequence has started as intended. The 53 second call ends as ST23 is about 2.7km/.17mi from the passing loop, putting it 4.5km/2.8mi from the crossing Mister Meintanis was concerned with. Mister Kennedy, having left a 115kph/71mph speed restriction at Heathcote Junction, begins to increase the train’s speed to the regular 130kph/81mph speed limit for the main line at Wallan. It doesn’t even reach that speed before, at 7:43pm, the data logger records an emergency stop being triggered. ST23 is approximately 100m/328ft from the points leading into the passing loop as the train’s brakes clamp down, but it’s too late. ST23 reaches the points leading into the passing loop at no less than 114kph/71mph and proceeds to derail almost immediately as it has no chance to follow the left-right turn the rails are forming beneath it. The leading locomotive falls over on its side as it cuts a path through the trackside trees, the five passenger cars derail behind it but don’t fall over entirely. Lastly, slowed by the derailment and resistance from the trees, the rear locomotive comes to a halt on track just beyond the points, largely undamaged.

A TV-crew captures this image of the derailed train the morning after the accident.

Aftermath

The crew supervisor used their handheld radio to issue a crew-wide radio call as soon as the train came to a rest, receiving replies from the whole passenger service crew but not getting a reply from Mister Kennedy (Mister Meintanis had no crew radio of his own). Two crew members independently used their personal radios to broadcast emergency calls seeking assistance and requesting that rail traffic be halted. Various passengers also used cellphones to call the emergency services, with the first such call being logged at 7:45pm. Calls also came in around the same time from the traffic control centers of both the regular gauge and broad gauge rail companies, giving further information to emergency responders. Passengers started to evacuate the stranded train before train crew members instructed them too (and thus before there was confirmation that traffic had been halted), and while they were told to stay off the rail line and leave luggage behind numerous passengers failed to follow at least one if not both of those instructions.

Two crew members, still failing to establish communications with Mister Kennedy, climbed up what used to be the underside of the locomotive and entered it through the right hand door, finding the locomotive crew trapped under gravel ballast and soil which had entered the cab during the derailment. They attempted to break the train’s windshield to facilitate ground-level entrance and possibly rescue the two men inside, but were unsuccessful. The first responder to arrive on site was a police officer who reached the wreckage at 8:03pm, soon followed by ambulances and firefighters, but they came too late to save the lives of the locomotive crew, pronouncing both men dead at the site. No one else died aboard the train, but 66 people were injured, 8 of which severely.

XP2018 sitting in the wreckage the day after the accident, with a ladder used by responders to access the door.

Investigators started by examining the signaling system and the points just behind the derailed train, finding no issue with either. The signaling system had not been in operation at the time of the accident, which was the reason for the individual permission (TA 17) and the notice about the passing loop, and the points were set and locked to the passing loop as intended. The investigation found that Mister Meintanis was employed by a labor-hire organization and worked for the rail company on a limited-time contract, with the fatal journey being his first time in the role of an AQW. He had the necessary training, just no practical experience on that job, having previously worked in other roles. He had reported to Donnybrook at the start of his shift, being briefed by an experienced AQW and receiving a refresher on his tasks, before travelling to KME16 where he would board ST23.

Mister Kennedy had around 40 years of experience, driving trains for most of his career before switching to a management-role, only returning to driving trains himself 4 years before the accident. He was fully licensed for driving the train on the line and there was no evidence of distraction, intoxication or a medical emergency affecting his performance on approach to the Wallan passing loop. The delays suffered along the train’s journey would have put the arrival in Melbourne beyond the end of his shift, but there was no reason to assume that he was significantly fatigued at the time of the accident.

It was never determined if he triggered the emergency stop as he realized the setting of the points just ahead or if Mister Meintanis, who wasn’t meant to actually interfere with the driving of a train, did.

The derailed train as shown in the report, photographed the morning after the accident.

As there was no voice recorder fitted to the locomotive (there was no requirement to have it) and with both locomotive crew members dying in the accident the investigation was forced to work with information from colleagues and knowledge from their experience, threading the needle of conducting a full investigation without jumping to conclusions, especially considering that an individual who might be at fault could no longer defend himself.

Mister Kennedy had been driving XPT services past Wallan 8 times in the prior 12 days, being kept on the main track every single time. The situation on southbound journeys had been largely identical each time, stopping at KME16 to let an AQW board. He would receive the permission to go from KME16 to Donnybrook (TA 17) and a warning-note regarding the level crossing (CAN 7) along with a Traffic Notice (TN) telling him that his train would run on the main track with the signals disabled and the points at the passing loop locked in the “straight ahead” position. He would then depart at regular speed while the AQW called the LCK as the XPT thundered past the start of the passing loop on a long straight piece of track where it could easily run at the 130kph/81mph speed limit. Mister Kennedy thus likely had a strong expectation of taking yet another trip past Wallan on the main track when he pulled up to KME16 ahead of the accident. He was once again handed three documents, 2 of which (TA 17 and CAN 7) were 100% identical with their copies he’d received on prior trips. Mister Meintanis had to hand him the notice about the passing loop and the associated speed limit (TN 367), but guidelines didn’t require the driver to confirm the instruction to the traffic controller. His omission of the changed routing through the loop from the radio call 11 minutes before the derailment indicates that he may have failed to notice the changed wording of the notice, since he received a near-identical printout on each prior trip, all of which told him that his train would run on the main track, along with two identical documents also concerning the section between KME16 and Donnybrook.

Human perception is not 100% factual, expectations based on past experience strongly influence the perception and make people more likely to see/hear what they expect to see/hear, even if that differs to a certain degree from what is actually there to be seen/heard. In simplified terms, Mister Kennedy may have received the two identical documents and then subconsciously registered “just another note” telling him about the main track at Wallan passing loop and acted accordingly, failing to notice that the wording of the note was different this time. A small error which turned fatal when the points came into view, too late to sufficiently slow down the train.

The forward part of the wreckage as shown in the report, with Car B sitting atop the mangled track.

The investigation also considered the possibility that Mister Kennedy lost track of his location after departing KME16 or that he erroneously looked at the signal for the parallel broad gauge line, but with no evidence to support either of those theories the report ends up back at the original theory. He didn’t seem confused about his position when stopped at KME16, and he also obeyed the changing speed limits between KME16 and the site of the accident. It’s also unlikely that he looked at the broad gauge tracks’ signals, since he would know that his signals weren’t operational. In conclusion, Mister Kennedy likely fell victim to a phenomenon in how humans perceive their surroundings, driven by his experience with the site, which made him act as if he was going to once again use the main track because as far as he knew, as he approached the site, he would use the main track. He didn’t forget something, or act out of malice or negligence, he, in simplified terms, trusted what he had seen, except what he had seen wasn’t what he had been shown.

With the cause of the derailment determined as certain as they reasonably could the investigation turned to the events once ST23 derailed, trying to find the reasoning for why the derailment had turned fatal despite the train retaining structural integrity. Workers cut down a number of trees once the on-site investigation had concluded, allowing the derailed train to be recovered and hauled to a maintenance facility by truck. XP2018 was recovered in one piece, except for the windshield which had been removed by responders to recover the bodies of the locomotive crew. The investigators noted that deformation to the body of the locomotive was very minor, but much of its left hand side had had the paint sanded down to metal during the derailment, indicating a significant distance traveled while already fully on its side.

XP2018 at a maintenance facility, showing bare metal on its left hand side.

Both left hand doors had been dislodged/separated from their frames during the accident, with the inward-opening driver’s cab door completely separating and ending up loose within the cab. This left a sizable opening through which significant amounts of dirt and gravel entered the cab, with the rear side of the door frame acing somewhat like a scoop as the train slid along. An examination of the hinges showed that the upper hinge of the forward left door had unfurled from around the frame-side hinge pin, while the lower hinge had torn off all six rivets holding it attached to the door frame.

The damaged door hinges as shown in the report. The upper hinge (2 images on the left) unfurled while the lower hinge (2 images on the right) separated from the frame in one piece.

The only strength requirement for the cab door, its hinge and locking mechanism was set at 2.5 Kilopascal (kPa) of inward pressure (equivalent to 0.025 bar/0.36psi), accounting for aerodynamic pressures when travelling at up to 200kph/124mph. There was no minimum pressure associated with the locomotive resting on its side, with guidelines only requiring that it be ensured the weight of the locomotive wouldn’t deform the frame enough to spring the door open. In fairness, the door frame hadn’t deformed to the point of popping the door open in the accident, but the total hinge failure had about the same result, if not worse as the door itself became part of the debris shoved into the cab. Investigators contracted an engineering firm, which calculated that XP2018 resting on its side at standstill exerted approximately 11 kPa (0.1 bar/1.5psi) on the door. They found that the weight of the resting train would exert enough force on the door to start deforming the hinges in a way similar to the damage suffered by the upper hinge. The forces suffered by XP2018’s door were assumed to be significantly higher though, with it just not resting on its side but falling onto it at considerable speed before dragging the damaged side along the ground. The investigators thus concluded that the door was not sufficiently engineered to withstand an overturning of the locomotive.

The investigation drew comparisons to a 2004 accident at Ufton Nervet (England) where a British InterCity 125 (on which the XPT is based) struck a car at a level crossing and subsequently derailed. The accident (which was previously covered on this blog [be aware that the article deals with a suicide]) also saw the death of both locomotive crew members and also involved the failure of the left hand driver’s cab door allowing ingress of dirt and gravel once the power car had fallen over at similar speed to XP2018. The report by the British investigators had pointed to an already ongoing research project to improve driver safety, feeding information from the accident to the researchers. That research-project had ended with the recommendation to install physical separations between the cab door and the driver’s area (or possibly moving the door elsewhere along the locomotive) along with changing the orientation of the door from being front-hinged to rear-hinged so a simple opening of it would see the driver shielded from the debris. While the opening of a still-hinged door was irrelevant to the accident at Wallan there was no adoption of a separation between the door and driver’s area either.

The investigators concluded that the forces suffered by the door as XPT2018 fell over went way beyond what it was both required to and capable of withstanding, tearing it off the hinges and allowing both the door itself and significant amounts of debris to enter the driver’s cab, which likely caused further injury to the possibly already injured locomotive crew, contributing to their perishing in the accident.

A British HST motor car (top left) compared to the XPT power car (top right) and the power cars from the Ufton Nervet (bottom left) and Wallan derailment (bottom right). The orange circles point to holes torn in the locomotive skin, something not suffered by XP2018.

The investigation ended up listing Mister Kennedy’s failure to obey the passing loop’s speed limit as the accident’s main cause, leading to his train entering the siding at approximately 8 times the speed the points are designed to handle. However, the investigation also points out that Mister Kennedy’s situation was far from ideal, with the change of operations being relatively spontaneous and there being no appropriate system to ensure drivers were aware of the changed routing as well as the lowered speed limit. Simply requiring drivers to tell traffic controllers of the speed limit instruction may have ensured that Mister Kennedy noticed the changed speed limit, especially with the signaling system being out of operation and thus unable to indicate that the points were set to direct the train into the siding. The AQWs sent to ride along in order to ensure operation of the nearby level crossing also didn’t offer the level of backup safety a secondary driver may have provided, yet no requirement for a second driver to accompany a train at the site had been put in place. The AQWs were seen as sufficient to hand over some paperwork and call the LCK positioned further down the line. This left a literal piece of paper as the only measure to ensure train drivers knew they were meant to break an established routine in a way that required preparation (in this case, deceleration) well before reaching the site, or rather, the possibility that the driver notices changed wording on a piece of paper he’s handed every single time he passes the site. The points themselves also garnered criticism, with the investigation questioning why points with a 15kph/9mph speed limit were used on a section with a 130kph/81mph speed limit when versions capable of handling higher speeds exist. These contributing factors are why the accident ends up being referred to as caused by “management failure” rather than simple “human error”.

Trees were cut down and a gravel road was constructed in preparation of the wreckage’s removal.

The report closes with 15 safety recommendations being issued, largely centering around the risk assessment of temporary traffic situations like the one encountered at Wallan and the improvement of how drivers receive safety critical information. Passenger service crew training was also advised to be improved, focusing on crowd control after incidents to avoid further endangerment in evacuations by passengers trying to bring luggage with them and/or evacuating a broken down train onto possibly non-blocked tracks. The former part is a known problem in the transportation industry, although it’s usually about passengers trying to bring their luggage when evacuating aircraft.

Lastly, the Australian Rail Industry Safety and Standards Board (RISSB) announced that they would review structural requirements for locomotive designs as well as see to improving emergency access, specially when a locomotive overturns, to facilitate faster rescue/evacuation of locomotive crews. Crew members and responders at Wallan had had to scale the (former) underside of the locomotive and then climb down into the cab until firefighters managed to remove the windshield (which marginally improved access, since its opening was fairly small), making it difficult to access people in the cab, not to mention trying to remove an injured/incapacitated person from the cab. The only other entrance was the rear right hand door, but the report explains that climbing through the length of the engine compartment wouldn’t have been practical either.

XP2018 was repaired after the investigation concluded and was renumbered XP2019, retiring its old number out of respect for the victims. It reentered regular service on the 24th of February 2023, showing no sign of its past. XP2000, the rear motor car involved in the accident, had already returned to service in 2021, retaining its original number.

XP2019, the repaired leading power car from the accident, captured by an enthusiast two days after reentering service.

_______________________________________________________________

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.

--

--

Max S

Train crash reports and analysis, published weekly.