Could RAIB's reliance on the On-Train Monitoring Recorder (OTMR) data for the impact speed be wrong?

Technologist · 24 Jun 2026

Moderator note: Split from

Report of collision Bedford (Elstow) - 19/06/26

Hearing reports of a collision between two trains at Bedford with one train now derailed

www.railforums.co.uk

Signal Head said:
The report says it was standing next to signal WH152 (next one ahead of 154).

To have accelerated to 76mph having passed a Yellow aspect is strange, to put it mildly.

Also noting the estimated collision speed as 49mph, the Hitachi unit has been pushed over 3 carriage lengths beyond the signal at which it was standing, not exactly a low speed collision as has been suggested in some quarters from the damage sustained.

For comparison, Clapham was estimated at 35mph, the fact that the vehicles here are mostly structurally intact is testament to the huge steps forward that have been made in modern rolling stock design.

I think the estimated collision speed of 49mph needs to be taken with a pinch of salt, Euro NCAP tests cars at 40mph into a semi deformable barrier and they lose 10-20% of their length doing so.

As a gross approximation hitting an object of similar mass that is not restrained and has similar energy absorbing capabilities to the impactor is about the same as hitting something solid at half the speed.

The impact attenuating nose of a high speed train is generally designed to resist a full loaded train travelling at ~15kph (in the case of Eurostar from memory) striking an immovable object without compromising the drivers compartment. Which doesn't appear to have happened on the class 810. This would suggest that the impact velocity was around the 30kph range.

RailUK Forums

AlterEgo · 24 Jun 2026

Technologist said:
I think the estimated collision speed of 49mph needs to be taken with a pinch of salt, Euro NCAP tests cars at 40mph into a semi deformable barrier and they lose 10-20% of their length doing so.

As a gross approximation hitting an object of similar mass that is not restrained and has similar energy absorbing capabilities to the impactor is about the same as hitting something solid at half the speed.

The impact attenuating nose of a high speed train is generally designed to resist a full loaded train travelling at ~15kph (in the case of Eurostar from memory) striking an immovable object without compromising the drivers compartment. Which doesn't appear to have happened on the class 810. This would suggest that the impact velocity was around the 30kph range.

Sorry, no - trains behave entirely differently to cars in crashes. The 810 is a very heavy object and was pushed three carriage lengths forward while it was at a stand and with brakes applied. Nobody in the industry I know had the immediate reaction of this being a low speed collision at "30kph" - this was a hard hit and several people I spoke to all concurred this was something like 40 to 50mph given the visual evidence. We all expressed immediate concern for the driver of the 360; this was extremely likely to be an unsurvivable incident for the driver.

I don't think there is any basis in suggesting the rear OTMR's speed recorder was faulty or has been misread; 49mph seems to me, about right.

GB · 24 Jun 2026

Technologist said:
I think the estimated collision speed of 49mph needs to be taken with a pinch of salt,

Does it? The RAIB are the ones with the data. Why would they estimate 49mph if it wasn't?

Technologist · 24 Jun 2026

GB said:
Does it? The RAIB are the ones with the data. Why would they estimate 49mph if it wasn't?

What's the sampling rate of the telemetry, were the sensors destroyed in the impact and the last few seconds were not yet written?

As a basic sense check if you drove your car (an object designed to resist this impact) into a larger stationary car at 49mph what would the results look like? To survive that impact with minor injuries you would need to be strapped in and have an airbag.

If the train had been doing 49mph at the point of impact we would expect to see, substantially greater damage to both vehicles, a very large number of people killed as they almost instantly (around 2X the crumple zone) got accelerated to greater than 10ms-1 (about the same as a 5m drop) before crashing into hard interior fixtures.

sharpener · 24 Jun 2026

Technologist said:
What's the sampling rate of the telemetry, were the sensors destroyed in the impact and the last few seconds were not yet written?

Well the deceleration from 76 to 49 mph over 9 s accords well with the spec for emergency braking of this stock according to other threads.

So the data seems to be internally consistent over at least that time period which kind of rules out gross measurement inaccuracy. The report seems to be silent on matters over which there is doubt, so I think what has been published is quite reliable.

They do say some of it may be revised, in particular when they have completed the analysis of the OTDR from the front of 1H46 which was recovered later than the others.

Bald Rick · 24 Jun 2026

Technologist said:
What's the sampling rate of the telemetry, were the sensors destroyed in the impact and the last few seconds were not yet written

The sampling rate is usually measured in small fractions of a second, from the previous reports I’ve been involved in.

If the RAIB say 49mph, it was 49mph or very nearly so.

== Doublepost prevention - post automatically merged: 24 Jun 2026 ==

Technologist said:
If the train had been doing 49mph at the point of impact we would expect to see, substantially greater damage to both vehicles

I guess you will need to amend your views in this matter when the report comes out.

Nym · 24 Jun 2026

I suppose that the level of damage to the front end at Grayrigg wouldn't be able to be used as evidence to support that the RAIB knows what they're on about?

edwin_m · 25 Jun 2026

Nym said:
I suppose that the level of damage to the front end at Grayrigg wouldn't be able to be used as evidence to support that the RAIB knows what they're on about?

Can you clarify this? More or less damage than you might expect?

For Elstow, the RAIB will (or may have already) corellate the OTMR data with the times of occupying and clearing track circuits that will be recorded on the signalling data recorders. There are also separate OTMRs for each cab, which can be checked against each other. An error in one of these is highly unlikely but would also be found by checking against the others.

AlterEgo · 25 Jun 2026

Technologist said:
What's the sampling rate of the telemetry, were the sensors destroyed in the impact and the last few seconds were not yet written?

As a basic sense check if you drove your car (an object designed to resist this impact) into a larger stationary car at 49mph what would the results look like? To survive that impact with minor injuries you would need to be strapped in and have an airbag.

If the train had been doing 49mph at the point of impact we would expect to see, substantially greater damage to both vehicles, a very large number of people killed as they almost instantly (around 2X the crumple zone) got accelerated to greater than 10ms-1 (about the same as a 5m drop) before crashing into hard interior fixtures.

You need to look at more pictures of train crashes, to be honest.

jawr256 · 25 Jun 2026

The RAIB update does acknowledge the possibility that the collision speed could be refined as the investigation continues.

https://www.gov.uk/government/news/collision-between-two-passenger-trains-near-elstow

Some of this information may be refined as RAIB’s investigation progresses. RAIB continues to analyse evidence relating to the sequence of events, the performance of the signalling system and the speed at which the collision occurred, all of which remain under investigation.

edwin_m · 25 Jun 2026

The above is the 1994 crash testing of the prototype of the "crumple zones" as fitted to post-privatisation rolling stock, as presented by Carol Vorderman on the late lamented "Tomorrows World" programme. It doesn't mention the speed but RSSB project T118 (accessible only to registered industry users) notes 60km/h (about 37mph) for what is probably the same series of tests.

Technologist · 25 Jun 2026

Bald Rick said:
The sampling rate is usually measured in small fractions of a second, from the previous reports I’ve been involved in.

If the RAIB say 49mph, it was 49mph or very nearly so.

I guess you will need to amend your views in this matter when the report comes out.

I'm engineer, the worst possible bit of evidence is "if someone has said it it must be true" based on their reputation. I have no doubt that RAIB have a piece of telemetry that either ends at 49mph or a G meter reading that registers and impact when the rail vehicle was doing 49mph. I can think of plenty of ways that said data might be miss aligned. I'm sure the RAIB can as well but they aren't going to get any flack for releasing a number from telemetry that they have and then revising it later when they look at how the vehicle bodies deformed.

I am aware that a train does not behave as a rigid body but even if you ignore the rest of the train and were to model the impact as just the two coaches that made contact with each other then the amount of damage and injuries is not consistent with the claimed impact velocity.

To overly simplify if we take the two carriages that impact at 22ms-1 and the combined mass travels forward at 11ms-1. The crumple zone appears to have deformed around 1-1.5m, the crash g forces for a fully restrained passenger would be 4-6G. However only the passengers in a rear facing seat are close to being restrained, someone in a forward facing seat is hitting the one in front at about 15mph which is serious to minor injury. Someone in a table seat is having their midriff hit the table edge at about 11mph, the equivalent of being dropped over 1 m, before having their head pounded into the table at ~25mph. Someone standing up is hitting the next hard surface at 20-25mph most likely head first. Those are under estimates because obviously the two end cars are not unrestrained. The fact that of the hundreds of people on board we only got 9 critical injuries and no deaths suggests that the impact forces were lower than this.

The carriage structures are not magic, they are made out of normal structural materials and the two vehicles have similar densities to a typical car. The crush distance and the lack of any crumpling on the other carriages or very much jack knifing also suggest lower velocities.

== Doublepost prevention - post automatically merged: 25 Jun 2026 ==

Nym said:
I suppose that the level of damage to the front end at Grayrigg wouldn't be able to be used as evidence to support that the RAIB knows what they're on about?

Completely different impact the carriages essentially dragged along the ground. Rail vehicles and cars crash differently because of the scale and the environment. As the rail vehicle is so much larger the world is relatively "smoother" and thus it will tend to loose velocity slower producing lower accelerations and less tumbling. They also tend to be larger and stronger than lots of objects in their environment, see car vs lamp post vs train vs OHLE suport.

However when they impact something truly immovable and perpendicular they tend to suffer more than a road vehicle at equivilent velocity as the passengers are less well restrained at the object has less strength to weight due to the square cubes law.

AlterEgo · 25 Jun 2026

Technologist said:
I'm engineer, the worst possible bit of evidence is "if someone has said it it must be true" based on their reputation. I have no doubt that RAIB have a piece of telemetry that either ends at 49mph or a G meter reading that registers and impact when the rail vehicle was doing 49mph. I can think of plenty of ways that said data might be miss aligned. I'm sure the RAIB can as well but they aren't going to get any flack for releasing a number from telemetry that they have and then revising it later when they look at how the vehicle bodies deformed.

I am aware that a train does not behave as a rigid body but even if you ignore the rest of the train and were to model the impact as just the two coaches that made contact with each other then the amount of damage and injuries is not consistent with the claimed impact velocity.

To overly simplify if we take the two carriages that impact at 22ms-1 and the combined mass travels forward at 11ms-1. The crumple zone appears to have deformed around 1-1.5m, the crash g forces for a fully restrained passenger would be 4-6G. However only the passengers in a rear facing seat are close to being restrained, someone in a forward facing seat is hitting the one in front at about 15mph which is serious to minor injury. Someone in a table seat is having their midriff hit the table edge at about 11mph, the equivalent of being dropped over 1 m, before having their head pounded into the table at ~25mph. Someone standing up is hitting the next hard surface at 20-25mph most likely head first. Those are under estimates because obviously the two end cars are not unrestrained. The fact that of the hundreds of people on board we only got 9 critical injuries and no deaths suggests that the impact forces were lower than this.

The carriage structures are not magic, they are made out of normal structural materials and the two vehicles have similar densities to a typical car. The crush distance and the lack of any crumpling on the other carriages or very much jack knifing also suggest lower velocities.

== Doublepost prevention - post automatically merged: 25 Jun 2026 ==

Completely different impact the carriages essentially dragged along the ground. Rail vehicles and cars crash differently because of the scale and the environment. As the rail vehicle is so much larger the world is relatively "smoother" and thus it will tend to loose velocity slower producing lower accelerations and less tumbling. They also tend to be larger and stronger than lots of objects in their environment, see car vs lamp post vs train vs OHLE suport.

However when they impact something truly immovable and perpendicular they tend to suffer more than a road vehicle at equivilent velocity as the passengers are less well restrained at the object has less strength to weight due to the square cubes law.

How do you explain a Class 810 - which likely had its brakes on - being shunted 75-80 metres? Does that seem like a 20-25mph collision to you?

The 360 did not just stop upon impact, it continued forward for 75-80 metres. It didn't hit a wall or a solid buffer.

Technologist · 25 Jun 2026

edwin_m said:
The above is the 1994 crash testing of the prototype of the "crumple zones" as fitted to post-privatisation rolling stock, as presented by Carol Vorderman on the late lamented "Tomorrows World" programme. It doesn't mention the speed but RSSB project T118 (accessible only to registered industry users) notes 60km/h (about 37mph) for what is probably the same series of tests.

Very interesting:

I would compare the activation of the crumple zones on the following carriages here vs what we see at Bedford and infer that the impact was likely at a lower velocity. Obviously the designs of the trains aren't the same and we had what might be termed "incompatibility" with the nose of the 810.

== Doublepost prevention - post automatically merged: 25 Jun 2026 ==

AlterEgo said:
How do you explain a Class 810 - which likely had its brakes on - being shunted 75-80 metres? Does that seem like a 20-25mph collision to you?

The 360 did not just stop upon impact, it continued forward for 75-80 metres. It didn't hit a wall or a solid buffer.

Do we know precisely how the brakes on either train work especially in a faulted condition?

1: I don't know what pressure the 810's brakes apply when stationary on flat surface or even if the actual friction brakes were applied given it can re-gen.
2: The impact could have damaged or confused the braking system or it could even be designed to reduce accident loads by not fully applying following an impact (I suspect that this is unlikely).
3: It might have been proceeding at a very slow speed with zero brakes on

Fundamentally both trains stayed mostly on the rails and that is a low friction environment so even low velocities can make something move a long way.

AlterEgo · 25 Jun 2026

Technologist said:
Very interesting:

I would compare the activation of the crumple zones on the following carriages here vs what we see at Bedford and infer that the impact was likely at a lower velocity. Obviously the designs of the trains aren't the same and we had what might be termed "incompatibility" with the nose of the 810.

== Doublepost prevention - post automatically merged: 25 Jun 2026 ==

Do we know precisely how the brakes on either train work especially in a faulted condition?

1: I don't know what pressure the 810's brakes apply when stationary on flat surface or even if the actual friction brakes were applied given it can re-gen.
2: The impact could have damaged or confused the braking system or it could even be designed to reduce accident loads by not fully applying following an impact (I suspect that this is unlikely).
3: It might have been proceeding at a very slow speed with zero brakes on

Fundamentally both trains stayed mostly on the rails and that is a low friction environment so even low velocities can make something move a long way.

A 25mph crash does not shift a Class 810 75-80m down the track. The 810 was standing with an AWS fault as per RAIB, and more likely than not had its brakes engaged.

I'd be interested to hear from any rail professionals who would agree with your estimate, which puts the crash at half the speed RAIB initially have cause to believe it to have occurred at.

This crash happened at 29mph on legacy stock with lower crashworthiness: https://en.wikipedia.org/wiki/Kirkby_train_crash

== Doublepost prevention - post automatically merged: 25 Jun 2026 ==

Technologist said:
I'm engineer, the worst possible bit of evidence is "if someone has said it it must be true" based on their reputation. I have no doubt that RAIB have a piece of telemetry that either ends at 49mph or a G meter reading that registers and impact when the rail vehicle was doing 49mph. I can think of plenty of ways that said data might be miss aligned. I'm sure the RAIB can as well but they aren't going to get any flack for releasing a number from telemetry that they have and then revising it later when they look at how the vehicle bodies deformed.

They have the forward facing CCTV from the 360 and have viewed it; it is very easy to work out speeds from video and RAIB said they will "refine" the speed of collision, which to me means within a small margin of error.

sharpener · 25 Jun 2026

Technologist said:
I'm engineer, the worst possible bit of evidence is "if someone has said it it must be true" based on their reputation. I have no doubt that RAIB have a piece of telemetry that either ends at 49mph or a G meter reading that registers and impact when the rail vehicle was doing 49mph...

I am an engineer too. Various ppl on here have modelled the crash in various ways (deceleration, conservation of momentum, loss of energy), some of it was manifestly wrong but the rest is consistent and agrees more or less with the RAIB figures.

RAIB will have access to much more historical data, much better models, and information from the two trains and marks on the track we are not privy to, so can assess the collision dynamics much better than we can here. I think until they publish any updates (which I would not expect to be major) we have to go with their present estimate of 49mph.

In any event it will not shed much light on the two burning questions which are why did 1B76 stop and why did 1H46 not.

Harpo · 25 Jun 2026

Other factors to consider are the initial speed of the 360 at the point where the brakes were applied, the known braking rate of the vehicles and the elapsed time to the collision. This may also have validated the impact speed.

Paul AC · 25 Jun 2026

Technologist said:
I'm engineer, the worst possible bit of evidence is "if someone has said it it must be true" based on their reputation. I have no doubt that RAIB have a piece of telemetry that either ends at 49mph or a G meter reading that registers and impact when the rail vehicle was doing 49mph. I can think of plenty of ways that said data might be miss aligned. I'm sure the RAIB can as well but they aren't going to get any flack for releasing a number from telemetry that they have and then revising it later when they look at how the vehicle bodies deformed.

I'm also an engineer. I'm an electronics engineer who has been designing electronic equipment for trains for nearly 40 years - and yes I do mean designing, not just using other people's boxes of electronics. I've designed diagnostic equipment for trains, including remote access and data collection, and also installed data recorders on trains.

The colliding unit has two data recorders, one for each driving cab. RAIB have told us that they have only, so far, been able to read the data from the "rear cab" data recorder, but that still records a lot of useful data such as speed and air brake pressures; it is mainly missing the driver activities in the leading cab such as response to AWS cautions, and also the passage of active TPWS loops.

It's very easy to take a sample of speed and time data from the data recorder and match these against the signalling system records of when the train passed certain junctions and signals; and of course the distances between them are known. It's then a simple maths task to check they all make sense, and so check the calibration of the data recorder's speed recording signal. I'm sure RAIB will have done this before they reported the 49 mph collision speed. It might be 1 or 2 mph out, but that's not important. It won't be half or twice what RAIB say it is.

Helvellyn · 25 Jun 2026

Firstly, I know many who consider that the first Desiros (360s/450s/444s/350s) were in several regards over-engineered by Siemens. They are heavy, well put together units.

Secondly, the images from the crash site didn't suggest a low speed collision. As well as the severe cab damage/deformation there were images showing that the first and second coaches of the 360 were crushed together - coach ends touching, visible end damage, corridor connections not visible, couplings probably sheered off. But the anti-climb features look to have locked the coaches in place, as designed to stop jack-knifing or overriding.

Thirdly, an interior photo from one of the 360 coaches (air thick with dislodged dust) showed the force of the collision had ejected the seat cushions and metal seat backs out of many of the seat frames - the seat cushions don't just pop off on those seats, let alone the actual metal backs.

Ultimately I think many presume this was a low speed collision because both trains looked intact and upright. No jack-knifed carriages and that horrendous concertina effect of some previous collisions.

Plus until the interim report I don't think anyone was assuming the second train could have been going so fast. It's still hard to accept that a 49mph impact speed - and the 360 doing 76mph seven seconds earlier - means that the second train hadn't just gone through a Red signal but was accelerating despite two previous Single Yellows (as forward facing CCTV shows). The investigation will have to determine what failures may have happened but sadly it feels like one of those classic Swiss cheese causations where a number of factors aligned, with tragic results.

Nottingham59 · 25 Jun 2026

Technologist said:
I am aware that a train does not behave as a rigid body but even if you ignore the rest of the train and were to model the impact as just the two coaches that made contact with each other then the amount of damage and injuries is not consistent with the claimed impact velocity.

The damage is substantial. The front cab of the 360 was destroyed and at several crumple zones between carriages have concertina'd. Railway vehicles are designed to withstand 2000kN (200 tons weight) of crushing force along their length.
How does that compare to your car?

The injuries were substantial. 100 people were taken to hospital, and as of today 38 are still there, 3 of them critically injured and one dead. As described here: https://www.bbc.co.uk/news/articles/c70y154rpnlo

How many people do you think would have been in the leading carriage of the 360 at the time?

Technologist said:
To overly simplify if we take the two carriages that impact at 22ms-1 and the combined mass travels forward at 11ms-1. The crumple zone appears to have deformed around 1-1.5m, the crash g forces for a fully restrained passenger would be 4-6G.

It looks to me as if 360 impaled itself at least 2.0m beyond the point where the coupler of the 810 (just inside the nose cover) and that of the 360 (below and well forward of the windscreen) would have made first contact. So more gradual than you calculated. Perhaps 3G for a rear-facing passenger? And every 1.0m of crumple zone in either train would have increased the crush distance experienced by the cars away from the two colliding, reducing the shock impact further.

Technologist said:
However only the passengers in a rear facing seat are close to being restrained,

That's half of them, then.

Technologist said:
someone in a forward facing seat is hitting the one in front at about 15mph which is serious to minor injury.

Maybe in the front carriage of the 360. RSSB standard: GMRT2100 Rail Vehicle Structures and Passive Safety how a typical impact might appear in this figure here:

As the unrestrained body flies forward towards the rapidly decelerating seat back in front of it, first contact will be the knees at relatively slow speeds (having travelled only 10cm or so), and this contact will then transmit force through the legs and pelvis slowing the lower torso subtantially before the head (which has travelled about 100cm) hits the seat back, at a point half-way through the two colliding vehicles interpenetrating each other, and therefore at a relative speed of 10-15 mph. That's easily survivable: skulls are incredibly strong against crush injury and the seat backs are design to minimise damage to people hitting them.
Passengers in the following carriages would have been protected to some extent by the effect of inter-carriage crumple zones. An extra 1m of crumple means the head hits the seat in front after one-third of the collision has happened, so a speed of 5-10mph.

Technologist said:
Someone in a table seat is having their midriff hit the table edge at about 11mph, the equivalent of being dropped over 1 m, before having their head pounded into the table at ~25mph.

Nah. The table edge is only 10cm from the belly. It will make contact after 1/20 of the collision has occurred. Much less that your 11mph relative speed. And the head will move around a metre, so the same as the airline seating example above.

Technologist said:
Someone standing up is hitting the next hard surface at 20-25mph most likely head first.

How many are likely to have been standing up?

The RSSB standard is free to read but you have to register for access:

GMRT2100 Iss 6.2 - Rail Vehicle Structures and Passive Safety

This document sets out requirements for the design and integrity of primary and secondary rail vehicle structures, including interior crashworthiness.

www.rssb.co.uk

Harpo · 25 Jun 2026

Huge thanks to those with technical insight for everything in this thread. The railway has a fantastic tradition of mutual improvement and it really shines here.

Thanks to, to @Technologist. There’s nothing wrong in not wanting to take things at face value.

Bald Rick · 25 Jun 2026

Technologist said:
I'm engineer, the worst possible bit of evidence is "if someone has said it it must be true" based on their reputation.

Indeed.

Given that the RAIB inspectors will have:

- Downloaded the speed data from the OTMR of the rear cab of the 360
- viewed the Forward Facing CCTV video, with clear visual references to the collision speed
- reviewed the signalling data from the trackcircuits (measured to the thousandths of a second)
- reviewed GPS data from the trains concerned
- measured the deformation of the carriages
- obtained detailed designs of the carriages from their owners and been able to conduct initial calculations
- measured the precise distances the trains have travelled from the point of collision
(and more)

… I would say that the RAIB have sufficient evidence to back up their statement that the collision speed was around 49mph.

Certainly more evidence than an engineer on a public forum who has none of the above evidence.

I know who I trust.

Technologist · 25 Jun 2026

Nottingham59 said:
The damage is substantial. The front cab of the 360 was destroyed and at several crumple zones between carriages have concertina'd. Railway vehicles are designed to withstand 2000kN (200 tons weight) of crushing force along their length.
How does that compare to your car?

The injuries were substantial. 100 people were taken to hospital, and as of today 38 are still there, 3 of them critically injured and one dead. As described here: https://www.bbc.co.uk/news/articles/c70y154rpnlo

How many people do you think would have been in the leading carriage of the 360 at the time?

It looks to me as if 360 impaled itself at least 2.0m beyond the point where the coupler of the 810 (just inside the nose cover) and that of the 360 (below and well forward of the windscreen) would have made first contact. So more gradual than you calculated. Perhaps 3G for a rear-facing passenger? And every 1.0m of crumple zone in either train would have increased the crush distance experienced by the cars away from the two colliding, reducing the shock impact further.

That's half of them, then.

Maybe in the front carriage of the 360. RSSB standard: GMRT2100 Rail Vehicle Structures and Passive Safety how a typical impact might appear in this figure here:

View attachment 206677
As the unrestrained body flies forward towards the rapidly decelerating seat back in front of it, first contact will be the knees at relatively slow speeds (having travelled only 10cm or so), and this contact will then transmit force through the legs and pelvis slowing the lower torso subtantially before the head (which has travelled about 100cm) hits the seat back, at a point half-way through the two colliding vehicles interpenetrating each other, and therefore at a relative speed of 10-15 mph. That's easily survivable: skulls are incredibly strong against crush injury and the seat backs are design to minimise damage to people hitting them.
Passengers in the following carriages would have been protected to some extent by the effect of inter-carriage crumple zones. An extra 1m of crumple means the head hits the seat in front after one-third of the collision has happened, so a speed of 5-10mph.

Nah. The table edge is only 10cm from the belly. It will make contact after 1/20 of the collision has occurred. Much less that your 11mph relative speed. And the head will move around a metre, so the same as the airline seating example above.

How many are likely to have been standing up?

The RSSB standard is free to read but you have to register for access:

GMRT2100 Iss 6.2 - Rail Vehicle Structures and Passive Safety

This document sets out requirements for the design and integrity of primary and secondary rail vehicle structures, including interior crashworthiness.

www.rssb.co.uk

I will probably read that, I was coming at this from a perspective of if you crashed a car at 49mph into a stationary car without wearing a seatbelt you would be quite likely to suffer fatal injuries and the vehicles would be mashed.

However I think the takeaways from here are:

The crumple zones are physically larger even if being a much smaller proportion of the length of the vehicle
The train actually behaves more like a motorway pile up than a single vehicle, the impact of subsequent vehicles is taken up by the inter carriage crumple zones so the impact is progressively less in carriages further away from the impact
The braking time is ~ 9 seconds or so which would give someone of normal physical capabilities time to sit down and/or brace, potentially people at the extreme ends of the train might have actually heard the impact a substantial fraction of a second before the G force actually occured.

In fairness I once crashed a TVR sideways into a tree at about 40mph and suffered no injuries and didn't even write the car off, crashes are unpredictable. The tree was elastic and the car hit at a point where it was very strong.

Dave W · 26 Jun 2026

Technologist said:
crashes are unpredictable.

Given your acknowledgement of this, could you expand on what you think is so glaringly wrong about RAIB's preliminary position that overrides the unpredictability of a virtually unprecedented situation?

Technologist · 26 Jun 2026

Dave W said:
Given your acknowledgement of this, could you expand on what you think is so glaringly wrong about RAIB's preliminary position that overrides the unpredictability of a virtually unprecedented situation?

I am acknowledging that the crash may well have been around 40mph (that crash test was at 37mph), see previous arguments as to why I was surprised by the lower number and severity of casualties given the high speed of the crash and the fact nobody is secured.

Whilst crashes are unpredictable if you have a trains worth of people on board you are going to get a reasonably predictable spread of outcomes.

Dave W · 26 Jun 2026

How significant is the 9mph difference between what you acknowledge and what RAIB say?

(I am not being difficult - I am not an engineer and am interested to understand more)

AlterEgo · 26 Jun 2026

Technologist said:
I am acknowledging that the crash may well have been around 40mph (that crash test was at 37mph), see previous arguments as to why I was surprised by the lower number and severity of casualties given the high speed of the crash and the fact nobody is secured.

Whilst crashes are unpredictable if you have a trains worth of people on board you are going to get a reasonably predictable spread of outcomes.

No you aren't necessarily going to get a predictable spread of outcomes. It very much depends on what actually happens. In this case the fact the trains remained upright with no ejections contributed to the zero passenger fatalities (so far).

hwl · 26 Jun 2026

Technologist said:
The crumple zones are physically larger even if being a much smaller proportion of the length of the vehicle

The train actually behaves more like a motorway pile up than a single vehicle, the impact of subsequent vehicles is taken up by the inter carriage crumple zones so the impact is progressively less in carriages further away from the impact

The crumple zones are also much stiffer overall than road vehciles and consist of several semi-parallel elements of differing stiffness.

The couplers aren't any where as rigid a people think the hidden part of the end of unit couplers (and most of the length of the intervehicle couplers) is a tube that is designed to crumple first (softest element) so the anti-climbers* (hidden behind the cosmetic fibre glass the 810 but visible on the 360) on the vehicle ends engage to turn the train(s) into a rigid unit that stays aligned.
*the red features in the TW video.

Modern crump zones are far stiffer than that in the TW video.

43096 · 26 Jun 2026

Technologist said:
The carriage structures are not magic, they are made out of normal structural materials and the two vehicles have similar densities to a typical car. The crush distance and the lack of any crumpling on the other carriages or very much jack knifing also suggest lower velocities.

I'd suggest you go and have a look at some of the photos. There is some very significant "crumpling" at the ends of the Class 360 vehicles involved - not just the leading cab. It's designed that way to keep the passenger areas as intact as they can - no broken windows and doors still openable - and the crumple zones have worked exactly as Siemens designed them.

Your comparison with a car is, frankly, utterly pointless. They are completely different vehicles with different standards. As an engineer, I though you would know this...

Magrar · 26 Jun 2026

How do we know the precise distance the stationary train moved?

Was there witness marks on the rails, for example? Precision GPS?

Could RAIB's reliance on the On-Train Monitoring Recorder (OTMR) data for the impact speed be wrong?

Member

RailUK Forums

Verified Rep - Wingin' It! Paul Lucas

Established Member

Member

Member

Veteran Member

Established Member

Veteran Member

Verified Rep - Wingin' It! Paul Lucas

Member

Veteran Member

Member

Verified Rep - Wingin' It! Paul Lucas

Member

Verified Rep - Wingin' It! Paul Lucas

Member

Established Member

Member

Established Member

Established Member

Established Member

Veteran Member

Member

Member

Member

Member

Verified Rep - Wingin' It! Paul Lucas

Established Member

On Moderation

On Moderation