Petteril Junction - Locked Axles and Freight.

Well the RAIB report for the Petteril Junction (Settle to Carlisle line) derailment last year is out, and it makes for interesting reading.

Petteril was one of a series of incidents involving severe flats and locked axles on freight rolling stock, including three instances at Llangennech, Pencoed and Ferryside between 2017 and 2021. While injuries have thankfully been avoided, the consequences of these range from two broken rails to a 446,000L fuel spill, a serious fire and the contamination of a marine & wetland Site of Special Scientific Interest (and as always, it could have been significantly worse). In all cases, the RAIB reports concluded that the incidents could have been avoided or significantly mitigated, and identified maintenance and operations practices improvements.

What Went Right

Petteril stands out from the others.

The wagons were in good condition, and well maintained. The derailed wagon specifically had completed its scheduled maintenance, and passed its annual brake test five months before. The axleboxes were all sound and the brake system, although simple, is modern and robust and was probably behaving as expected. All persons involved in this incident were experienced; the driver was coming up on four decades in industry and the locomotive data evidences their driving and braking was entirely normal and within stated acceptable limits for October in the north of England. The Settle to Carlisle line has frequent passenger and freight services and is maintained accordingly; a Railhead Treatment Train (RHTT) had been through 22 hours before the incident train and was due to pass through again ninety minutes later. In any case the section at Dent – where the flat formed – was not considered abnormally high risk for leaf fall issues. The weather was cold – close to the local dew point – but not rainy.

That might seem an exhaustive list but what struck me reading it was the lack of precursors - there's very little there which jumps out as an obvious cause. In short: the vehicles were in good condition prior to the incident; they were driven correctly; low adhesion was a known but apparently controlled issue. The risks were known and the organisations had made efforts with their assets under their control to mitigate these.

Despite this, the wagon formed a flat and then travelled for an hour with a locked axle, damaging points and ultimately derailing on a bridge over the River Petteril where two of the wagons ended up.

Changing Perceptions

Fundamentally this does seems to be about risk. Modern passenger rolling stock is almost universally fitted with wheel slide protection (WSP) systems designed to modify the tractive or braking forces on a wheelset when they slip in low-adhesion conditions. This is possible on passenger trains because power and data connections are available down the length of the train; the same systems have not been fitted to freight trains precisely because those requirements can only be met on locomotives, not on the wagons, and the cost of overcoming these issues is perceived as disproportionate to the risk of managing wheel flats and potential derailments. Attempting to manage these complex risks across different operational and asset-owning organisations is another challenge, and one discussed in the Llangennech report.

Both understanding of the risk and the barriers to entry seem to be changing. The report’s first recommendation calls for a joint review of “the risks faced by freight wagons during normal brake applications in foreseeably low adhesion conditions”, including a detailed assessment of the risk of self-sustaining flats which may lead to a derailment. Furthermore, it calls for this review to “also identify what mitigations may be necessary to ensure that these risks are adequately controlled”. With widespread expectations that rail freight’s market share will only increase, and the monetary and reputational cost of disruption increasingly high (to say nothing of the public interest in avoiding, like Llangennech, things like massive environmental pollution and the temporary evacuation of three hundred homes).?

Some of these mitigations and their pitfalls are discussed explicitly in the report. Impact detection systems pick up flats on rotating wheels but not on locked axles; hot axlebox detectors only target failed axle bearings at a late stage; hot wheel detectors have historically struggled with false positives; axle counters can detect sliding wheelsets but only in some instances. Furthermore, all of these are fitted at fixed locations across the network meaning sampling rates are lower. All of them come with not insignificant capital expenditures.

Solutions

As a vehicles engineer I’m naturally biased towards a vehicle-borne solution, and WSP (or some system like it) seems promising if the power and data challenges can be met or avoided. The best form of asset condition monitoring is continuous, after all, and that can’t be achieved with lineside systems.

Industry is increasingly considering such options – according to the report VTG had been trialling a possible solution at the time of the incident, although not on the train in question – and Infinitive has been taking an interest in them too. In particular we’re looking at options which sidestep some of the issues with conventional power supplies such as air-to-power systems or large batteries, mitigating the potential negative impact on the wagons’ safety case. The other key attraction of these systems is that (unlike lineside options) it’s within the gift of ROSCOs and FOCs to fit, monitor and respond to.

The report's worth a read (as usual for the RAIB). The problem isn't going away.