Scrutable

Horrible news coming from Scotland today. Appears that at least three people have died at Carmont near Stonehaven after a train derailed and fell down an embankment. Earlier it was suggested that one carriage was submerged in the river.

Only 9 people on board, 6 passengers and 3 crew?

The train has been well and truly f.cked up. Can’t tell in the images I’ve seen if there was any landslide. But the front three carriages are all kinds of damaged, including fire damage, and the back two have ridden up over. Must have been absolutely terrifying.

lpm wrote: Wed Aug 12, 2020 3:27 pm Only 9 people on board, 6 passengers and 3 crew?

Sounds like it, yes, though I'm not sure if that's been formally confirmed. If it was busier then it could have been much worse.

Grumble wrote: Wed Aug 12, 2020 5:02 pm The train has been well and truly f.cked up. Can’t tell in the images I’ve seen if there was any landslide. But the front three carriages are all kinds of damaged, including fire damage, and the back two have ridden up over. Must have been absolutely terrifying.

E0F5018E-0278-446B-9F67-E8DDCBB12578.jpeg

Yeah, the HSTs are warhorses in terms of service but they weren't built to withstand that kind of rolling damage in the way that pendos were.

Overall, it's the first fatal train crash on the network in thirteen and a half years, the first multi-fatality crash caused by the infrastructure in eighteen years, and the first landslip fatality in 25 years. A good run, but awful that it wasn't much, much longer.

Hopefully there'll be more interest in R&D on the civils side to shore up the impact of climate change on the railway. There was already a reasonable amount, but perhaps now they'll move faster.

El Pollo Diablo wrote: Thu Aug 13, 2020 7:54 am

lpm wrote: Wed Aug 12, 2020 3:27 pm Only 9 people on board, 6 passengers and 3 crew?

Sounds like it, yes, though I'm not sure if that's been formally confirmed. If it was busier then it could have been much worse.

A death rate of 33% is pretty awful.

A train derailed in Italy on the Milano-Lecco line with only one person on board... who wasn't any kind of train crew.

Not entirely clear why the train set off on its own or if they knew there was a 49-year-old passenger in the last carriage (who sustained only a bruise to the knee) when it was decided to change the points to derail the train 6 km later. Two train operators (driver and conductor) also seem to have been injured trying to run after the train at the station.

Very similar incidents have happened here, too. Unfortunately it's a lot more common than you'd expect from a miraculous and complex feat of engineering

Interesting that "3 Killed in Train Crash" features prominently in the national media but "3 Die in Car Crash" would be found at the bottom of page 9 of the local weekly.

JQH wrote: Thu Aug 20, 2020 12:57 pm Interesting that "3 Killed in Train Crash" features prominently in the national media but "3 Die in Car Crash" would be found at the bottom of page 9 of the local weekly.

You’re right, but it does deserve to be more noted than a car crash with an equivalent death toll because the potential toll was much higher. When there’s a motorway pile up it makes the news.

These kinds of accidents are, unfortunately, going to happen more often. A senior NR geotech engineer said that even soil cuttings in brand new condition would have struggled with that volume of rainfall, which is only going to happen more often. I'll be very surprised if we go another 13 years without another fatal train crash.

Climate change strikes again.

El Pollo Diablo wrote: Sun Aug 23, 2020 9:36 pm These kinds of accidents are, unfortunately, going to happen more often. A senior NR geotech engineer said that even soil cuttings in brand new condition would have struggled with that volume of rainfall, which is only going to happen more often. I'll be very surprised if we go another 13 years without another fatal train crash.

There are other countries with more intense rainfall, and their railways will be engineered accordingly. That geotech engineer needs to consider that the standards (e.g. drainage capacity) they're working to are out of date.

Doesn't really matter - Carmont didn't have any earthworks in brand new condition - the point was to outline the impact of climate change on old infrastructure engineered for a different climate. Replacing the structure and drainage of many thousands of km of soil cuttings and embankments in a short or even long space of time isn't going to happen.

El Pollo Diablo wrote: Mon Aug 24, 2020 11:21 am Doesn't really matter - Carmont didn't have any earthworks in brand new condition - the point was to outline the impact of climate change on old infrastructure engineered for a different climate. Replacing the structure and drainage of many thousands of km of soil cuttings and embankments in a short or even long space of time isn't going to happen.

Climate change adaptation is going to cost a fortune and it's going to be extremely complex and difficult. Maybe we'll be abandoning some lines long term for example. Honestly f.ck knows, it's potentially an almost impossible mess.

Start by considering a spherical cow Wouldn’t some sort of there-has-been-a-landslide detection system be cheaper in the short term?

nekomatic wrote: Mon Aug 24, 2020 11:23 pm Start by considering a spherical cow Wouldn’t some sort of there-has-been-a-landslide detection system be cheaper in the short term?

Would electrified rails respond to a load of wet muck being dumped on the track?

I'm not sure of the electrical conductivity of wet soil, but if it's in any way reasonable then track circuits would show an occupied section if the landslip covered both tracks. Obviously, third rail electrification would short out, but only in the specific electrical section where the landslip was. The problems with these ideas are that:

1. If the landslip only covers one track, neither track circuits nor third rail electrification would detect that, and it would still derail the train.
2. Third rail electrification is only installed in certain parts of the country (south east & Merseyside), is essentially obsolete and NR doesn't want to install any more.
3. Third rails also swap from side to side on a regular basis - if the third rail was in the middle of the track, it is less likely to detect a slip.
4. Track circuits are much more common on the rail network, but they aren't installed everywhere, and they're also seen as an old train detection technology which isn't as beneficial as the preferred one (axle counters).
5. There are plenty of areas, particularly in Scotland, where there is no active train detection at all, and the method of ascertaining whether there's a train in the section is whether you or an earlier train driver has possession of a stick.
6. At Carmont, this is unfortunately the case - it is a very small mechanical interlocking using semaphore signals, with just two track circuits near to the signal box. The crash happened around 1.5 miles from the box, where there is no train detection at all. An injured passenger (who was a rail worker) had to walk along the tracks back to the signal box to alert the signaller and stop the traffic.
7. The future of signalling systems is to have no lineside train detection at all - onboard GPS (or similar) is the ambition. This is a long way from universal installation but it would be rather poorly thought through to begin relying on a technology marked for removal to help out with landslips.

The 2018 Earthworks Technical Strategy is here, and talks about some of the technological solutions that could be employed. I haven't read it, but a short one-line summary was passed on, which is that the tech isn't really there for a whole railway yet. But, one imagines, this is likely to see some renewed effort to make something work. It's far easier to think of reasons why things can't be done than why they can, especially on a railway, but this needs to get solved asap.

Worth also remembering that we'd likely start off by identifying the high-risk areas, then the exact areas of interest, then seeing what tools we've already got that might help.

We don't need to do the whole railway, 99.9% of it will be fine. Of course, all that's needed for a derailment is a couple of metres of wonky track.

My thoughts were along the lines (sorry) of some sort of pipe or tube or something running alongside the track in high-risk locations, with a means to detect if it gets damaged or distorted such as would happen in the event of a landslip. Possibly lasers are involved. Costly, but maybe a lot less than stopping the land from ever slipping. Would need to resist false alarms from wild weather, heavy snow and so forth, but maybe an alarm doesn’t necessarily mean you have to stop all trains and send out a maintenance party, just you put in a speed restriction in the area so any incident is much less serious.

Or just webcams on posts. Lots of webcams on posts. And machine vision.

Those are my ill-informed ramblings, anyway.

plodder wrote: Tue Aug 25, 2020 3:17 pm Worth also remembering that we'd likely start off by identifying the high-risk areas, then the exact areas of interest, then seeing what tools we've already got that might help.

We don't need to do the whole railway, 99.9% of it will be fine. Of course, all that's needed for a derailment is a couple of metres of wonky track.

Yeah, there are ideas out there, and lots of research both commercial and university-led. Inclinometers and so on. It's not really my field but the risk-based rollout would certainly be how it's done. Anyone interested can have a look at the tech strategy up there and let me know what's in it

I missed off a couple of things from my list above, which is that track circuits would only work for cuttings, and not embankments, where a slip would remove the land underneath the rails, again possibly leading to a derailment (a slip like this, thankfully without incident, happened near to where I live, on the WCML, a few years ago).

El Pollo Diablo wrote: Wed Aug 26, 2020 9:52 am
I missed off a couple of things from my list above, which is that track circuits would only work for cuttings, and not embankments, where a slip would remove the land underneath the rails, again possibly leading to a derailment (a slip like this, thankfully without incident, happened near to where I live, on the WCML, a few years ago).

There was a slip like this near Hook not that long ago. Embankment washed out from under the tracks, again without incident

nekomatic wrote: Tue Aug 25, 2020 11:21 pm Or just webcams on posts. Lots of webcams on posts. And machine vision.

Or drones flying ahead of the trains. </Cummings>

I did actually think of that one, but guessed that top speed aside, the kind of weather where landslips happen might not be ideal for drones.

You just need something which runs on the rails ahead of the train by a bit more than the safe stopping distance.