GWR withdraw some 800's due to cracks (ORR Report now published)

[LNW-GW Joint]

I can't see any blame attached, all parties worked well together to find the root cause.
The product complied with the specification and relevant standards, but began to fail in real life conditions.
1247 class 80x vehicles will have to be repaired, plus 487 class 385/395 vehicles, over 6 years.
That sounds like a huge cost to the railway, but it does not seem that this will all lie with Hitachi.

I can't see anything about newer builds of AT200/300 (ie EMR's 810s and Avanti's 805/7s), and what impact this situation might have on their production and introduction.

 

[Bletchleyite]

Just listened again on BBC Sounds and heard reference to the ORR Report and mention of:
"Trains experiencing greater loads from movement than allowed for in the original design".

In short, specced for better track than we've got? Oops.

I wonder would the cost-benefit of bringing the track up to German standards (say) compare better overall than beefing them up to cope with our bad track? Obviously they'll need repairing even if back to original spec, though.

 

[GC class B1]

I wonder whether the higher loads experienced on the bolster may be related to the poor ride and consequent higher stresses on the yaw dampers.
As I recall the class 395 had a poor ride when first introduced and they operate on HS1 which has good quality track.

 

[Mag_seven]

8 Hitachi class 800 5 carriage trains have been withdrawn from service due to significant cracks found in the yaw damper bolsters. Not just CAF who appear to have problems with designing/manufacturing vehicles to cope with UK infrastructure.....

Thread reopened as ORR report has been published:

Learning the lessons: ORR review into Hitachi Class 800 series and Class 385 cracking issue | Office of Rail and Road

ORR immediately launched a review of both the passenger and safety impact following Hitachi Rail Class 800 series trains withdrawn from service for safety checks on Great Western Railway, London North Eastern Railway, TransPennine Express and Hull Trains, along with cracks being separately...

www.orr.gov.uk

Learning the lessons: ORR review into Hitachi Class 800 series and Class 385 cracking issue

ORR immediately launched a review of both the passenger and safety impact following Hitachi Rail Class 800 series trains withdrawn from service for safety checks on Great Western Railway, London North Eastern Railway, TransPennine Express and Hull Trains, along with cracks being separately discovered after examination on ScotRail’s Class 385 and Southeastern’s Class 395 trains in May 2021.

 

[brad465]

I wonder whether the higher loads experienced on the bolster may be related to the poor ride and consequent higher stresses on the yaw dampers.
As I recall the class 395 had a poor ride when first introduced and they operate on HS1 which has good quality track.

The BBC report seems to focus on salt corrosion as being responsible:

Hitachi train cracks report calls for standards review

Hitachi Class 800 and 385 series trains were temporarily taken out of service in May 2021.

www.bbc.co.uk

Industry standards should be reviewed after fatigue and corrosion led to high-speed trains being withdrawn from service, the rail regulator has found.
Cracks were found in more than 180 Hitachi trains being used by Great Western Railway (GWR) and London North Eastern Railway (LNER) last year.
The Office of Rail and Road (ORR) said the damage was partly due to the trains being corroded by salt in the air.
Hitachi Rail said it took "appropriate action to prioritise safety".
In May 2021, the withdrawal of the damaged Hitachi Class 800 and 385 series trains led to a week of delays and cancellations for travellers.
Affected operators were Great Western Railway, London North Eastern Railway, TransPennine Express and Hull Trains.
Cracks were also found after examination of ScotRail's Class 385 and Southeastern's Class 395 trains.

The ORR investigation found the design of the trains complied with existing standards but that corrosion cracks were found on lifting points - which allow carriages to be raised during maintenance work - resulting from the use of a particular type of aluminium, which was corroded by salt in the air.
It said fatigue cracking was caused by the "trains experiencing greater loads from train movement than allowed for in the original design".
It added: "It is not yet known for certain why this happened, although potential factors include wheel wear and track design."
The trains first entered service in 2017.

 

[swt_passenger]

The BBC report seems to focus on salt corrosion as being responsible:

Hitachi train cracks report calls for standards review

Hitachi Class 800 and 385 series trains were temporarily taken out of service in May 2021.

www.bbc.co.uk

I’ve only scanned the report so far, but I think the BBC have only really reported about what appears to be the lesser problem of the two?

 

[samulih]

I can't see any blame attached, all parties worked well together to find the root cause.

I think this is good thing, investigate and report, too much hiding these days.

 

[Class 170101]

As I recall the class 395 had a poor ride when first introduced and they operate on HS1 which has good quality track.

That was because a balance had to be struck between HS1 track and track on the classic network in terms of ride quality.

 

[coppercapped]

That was because a balance had to be struck between HS1 track and track on the classic network in terms of ride quality.

Pedant Alert...!

As I understand the issue - not quite!

The suspension damping had to enable stable running at 140mph on HS1 which essentially means that the wheelsets and bogie have to be 'stiff' in rotation about a vertical axis. However this is not a good choice when running on more curved track at lower speeds as rail wear on the gauge corner is increased over that which would occur with bogies with 'softer' settings.

There was also an issue with poorer ride in the tunnelled sections of HS1 at lower speeds.

To overcome, or at least ameliorate, the issues suspension bushes with changed stiffness characteristics were used, they are stiffer to high frequency movements than to lower frequency movements.

So in a sense the problem was to do with different track characteristics but mainly in the sense of the different speed regimes and curvatures between HS1 and the 'classic' network.

 

[Hunter7091]

This is perhaps a noob question, but...

Seeing as the cracks are a pretty serious safety issue, wouldn't it make more sense to immediately implement the permanent solution on all the units, even if it means significant service disruption for weeks or even months? Doing it over 6 years will minimise the disruption, but what if a similar problem happens in that period?

 

[Bletchleyite]

This is perhaps a noob question, but...

Seeing as the cracks are a pretty serious safety issue, wouldn't it make more sense to immediately implement the permanent solution on all the units, even if it means significant service disruption for weeks or even months? Doing it over 6 years will minimise the disruption, but what if a similar problem happens in that period?

Cracks in engineering "stuff" are not at all uncommon and are monitored carefully with calculations done to ensure they don't pose a risk at that point. To use an example, the old Forth Road Bridge is known to have cracks, but these are within tolerance for the lighter traffic now allowed over it. When they get too large, it will need to be either repaired or closed completely.

I thought like you did before I did the excellent module the Open University offers on it. This free course they offer might cover some of it from a quick Google:

Open Learning

www.open.edu

 

[Irascible]

This is perhaps a noob question, but...

Seeing as the cracks are a pretty serious safety issue, wouldn't it make more sense to immediately implement the permanent solution on all the units, even if it means significant service disruption for weeks or even months? Doing it over 6 years will minimise the disruption, but what if a similar problem happens in that period?

Aside from it being a calculable risk to manage ( as long as something else doesn't show up as you said ) it's not a small job!

It's also possible this repair will generate another issue itself. If the vehicles are to spec and the spec is bad despite having masses of detailed data of our network, then I'm still suspecting a flaw in simulation & if it's still there, then new calcs for the repaired structure are also going to be flawed. Unless some large margin of error was worked in, anyway ( let's hope ). Possibly the supplied data is bad, which ends up in the same place.

 

[Wilts Wanderer]

This is perhaps a noob question, but...

Seeing as the cracks are a pretty serious safety issue, wouldn't it make more sense to immediately implement the permanent solution on all the units, even if it means significant service disruption for weeks or even months? Doing it over 6 years will minimise the disruption, but what if a similar problem happens in that period?

There is a limited capacity to perform friction-stir welding of aluminium structures, it is an exceptionally specialised skill and requires very particular factory facilities, including climate / humidity control. (Not getting this right during manufacture is one of the reasons the welds have failed in service.)

Also bear in mind that the trains effectively have to be stripped right back to bodyshell - including removal of all electronics, which would be damaged by the weld repair process. The complexity of the process is why a special factory is planned (Eastleigh I believe?) to effect these repairs.

I wouldn't see 6 years as a slow-progress, its probably the quickest rate the entire repair programme can be achieved.

 

[Towers]

Can anybody offer any insight into whether the increasingly pronounced issue of the bodyshell metalwork seemingly reacting beneath the paint, causing the paint to flake, is apparent on all 80x fleets? This is now very noticeable around the vehicle ends, door and window apertures across the fleets being operated by GWR, both the early vinyl-covered 800s and the later builds. Does this bear any relation to the cracking problems?

 

[Clarence Yard]

No, it’s nothing to do with 7000 series aluminium degradation. It’s just coming off.

 

[Bryson]

It's also possible this repair will generate another issue itself. If the vehicles are to spec and the spec is bad despite having masses of detailed data of our network, then I'm still suspecting a flaw in simulation & if it's still there, then new calcs for the repaired structure are also going to be flawed. Unless some large margin of error was worked in, anyway ( let's hope ). Possibly the supplied data is bad, which ends up in the same place.

The gist of the report was that the forces acting on the bolster are greater than the model indicated when fed with the network data. However during the investigation real world measurements have been made with sensor equipped 80X trains. It is these true measurements that have informed the new design.

It remains an open question as to what was wrong in the first instance - the model or the network data. This is something that needs to be resolved.

 

[Towers]

No, it’s nothing to do with 7000 series aluminium degradation. It’s just coming off.

Poor paint?

 

[PG]

Poor paint?

IIRC Pendolinos have a warranty on their paint, maybe the 8** series do too in which case someone will have to rectify the defects.

 

[northernbelle]

No, it’s nothing to do with 7000 series aluminium degradation. It’s just coming off.

Unfortunately I don't think it's just a case of the paint coming off - there does some to be metal corrosion occurring in specific areas of the bodywork - the corners of the carriage ends at cant rail height and about halfway up the door pockets seem to be particular locations. The fact that vinyl is bubbling up on the 800/0s suggests what it's stuck to is reacting rather than the paint/vinyl.

 

[fgwrich]

Poor paint?

Unfortunately I don't think it's just a case of the paint coming off - there does some to be metal corrosion occurring in specific areas of the bodywork - the corners of the carriage ends at cant rail height and about halfway up the door pockets seem to be particular locations. The fact that vinyl is bubbling up on the 800/0s suggests what it's stuck to is reacting rather than the paint/vinyl.

This doesn't surprise me at all. The door design is one of the pitfalls with the 80X Design - had they bitten the bullet and gone for, and I quote Hitachi, "Unreliablie" "European style plug doors" then we wouldn't have been left with the pocket design which not only acts as a dirt trap but also something of a water trap (as well as having issues in it's own right).

I'm also not surprised about potential paint issues either - you only have to look at the door framework of the Pistoia built units to see how utterly terrible it is. It really is just slapped on with a paint brush. Hitachi may have done it just to get the insides of the door pockets green, but as highlighted above it may now be coming back to haunt them as a water trap. I'll dig out my photos later.

 

[Irascible]

It remains an open question as to what was wrong in the first instance - the model or the network data. This is something that needs to be resolved.

Presumably it's the cause of the problems with CAF stock too. Bombardier & Siemens must have had the same data, surely? ( although I guess Bombardier had a few decades of application to apply also ).

 

[fgwrich]

Picking up on the paint finish quote from a few posts ago, I noticed that the paint was rather flaking away on 800303 in a few areas on more than one vehicle tonight.

 

[C96]

That doesn’t look appealing…

 

[Ashley Hill]

Possibly caused by hitting vegetation.

 

[Snow1964]

The gist of the report was that the forces acting on the bolster are greater than the model indicated when fed with the network data. However during the investigation real world measurements have been made with sensor equipped 80X trains. It is these true measurements that have informed the new design.

It remains an open question as to what was wrong in the first instance - the model or the network data. This is something that needs to be resolved.

There is also a risk that the sensors are only as good as where they have been located. The forces transmit through the structure and in layman’s language bend or twist it if it is not strong and rigid.

If you start strengthening corner joints etc, you can make these stronger than other parts, and if not careful you move the problem. I always remember this being illustrated with a kitchen sponge, if you twist it, it bends all the way along, but if you clamp the ends so they are rigid and do same, then the movement is concentrated on a smaller area so that is more likely to tear now. Obviously same will now happen to the metal frame of the trains, if you strengthen one area and same forces are applied then the remainder gets more flexing force (force is concentrated over smaller length)

So clearly if the other parts were not specified with a large margin then there is more risk on the remainder of the frame. But we now know there wasn’t sufficient margin and parts cracked, and if no one has used sensors on a selectively strengthened frame, how can anyone know how the real world forces act on the remainder, if certain parts that were absorbing some of the force are now rigid and unable to absorb any movement.

 

[GC class B1]

There is also a risk that the sensors are only as good as where they have been located. The forces transmit through the structure and in layman’s language bend or twist it if it is not strong and rigid.

If you start strengthening corner joints etc, you can make these stronger than other parts, and if not careful you move the problem. I always remember this being illustrated with a kitchen sponge, if you twist it, it bends all the way along, but if you clamp the ends so they are rigid and do same, then the movement is concentrated on a smaller area so that is more likely to tear now. Obviously same will now happen to the metal frame of the trains, if you strengthen one area and same forces are applied then the remainder gets more flexing force (force is concentrated over smaller length)

So clearly if the other parts were not specified with a large margin then there is more risk on the remainder of the frame. But we now know there wasn’t sufficient margin and parts cracked, and if no one has used sensors on a selectively strengthened frame, how can anyone know how the real world forces act on the remainder, if certain parts that were absorbing some of the force are now rigid and unable to absorb any movement.

This is a good point. Many years ago I was involved in investigating the reasons for a large number of broken springs on Vale Of Rheidol coaches when the line was owned by BR. The original springs were breaking and been replaced with new springs of a thicker section and instead of solving the problem the redesigned springs were also breaking. The reason for the failure of the redesign was that the springs had higher stresses as a result of the reduced movement allowed by the larger thickness. This had not been realised when the new springs were manufactured. The solution was to use a different steel to pretty much the original dimension.

 

[Peter Sarf]

There is also a risk that the sensors are only as good as where they have been located. The forces transmit through the structure and in layman’s language bend or twist it if it is not strong and rigid.

If you start strengthening corner joints etc, you can make these stronger than other parts, and if not careful you move the problem. I always remember this being illustrated with a kitchen sponge, if you twist it, it bends all the way along, but if you clamp the ends so they are rigid and do same, then the movement is concentrated on a smaller area so that is more likely to tear now. Obviously same will now happen to the metal frame of the trains, if you strengthen one area and same forces are applied then the remainder gets more flexing force (force is concentrated over smaller length)

So clearly if the other parts were not specified with a large margin then there is more risk on the remainder of the frame. But we now know there wasn’t sufficient margin and parts cracked, and if no one has used sensors on a selectively strengthened frame, how can anyone know how the real world forces act on the remainder, if certain parts that were absorbing some of the force are now rigid and unable to absorb any movement.

Interesting thoughts. A case of moving the problem around and not addressing the basic cause. The least one can hope for is that the problem can be moved to components that are easily replaced. Now what would they be on a 80x ?.

This is a good point. Many years ago I was involved in investigating the reasons for a large number of broken springs on Vale Of Rheidol coaches when the line was owned by BR. The original springs were breaking and been replaced with new springs of a thicker section and instead of solving the problem the redesigned springs were also breaking. The reason for the failure of the redesign was that the springs had higher stresses as a result of the reduced movement allowed by the larger thickness. This had not been realised when the new springs were manufactured. The solution was to use a different steel to pretty much the original dimension.

Another twist. In this case could the stronger springs have pushed the problem onto less easily repaired/replaced components ?.

Of course, perish the thought, but maybe the permanent way needs to be kinder to the 80x ?. Are the culprits just a few rogue locations on the network ?. Not really the fault of the network of course and not a fault of the original spec it seems BUT maybe that is the cheaper solution overall ?.

 

[GC class B1]

Interesting thoughts. A case of moving the problem around and not addressing the basic cause. The least one can hope for is that the problem can be moved to components that are easily replaced. Now what would they be on a 80x ?.


Another twist. In this case could the stronger springs have pushed the problem onto less easily repaired/replaced components ?.

Of course, perish the thought, but maybe the permanent way needs to be kinder to the 80x ?. Are the culprits just a few rogue locations on the network ?. Not really the fault of the network of course and not a fault of the original spec it seems BUT maybe that is the cheaper solution overall ?.

The new springs on the V of R coaches were also breaking and that was the problem I was dealing with. From memory the thicker section of the spring material and I think it was square section and not the original round section meant it had a higher stiffness and this resulted in higher stresses but I can’t remember the exact mechanism for failure. The point was that while on the face of it increasing the metal thickness should reduce the risk of breaking, not fully understanding the mechanism of failure meant that the solution used didn’t solve the problem.

The new springs were probably designed before 1970s and technology has moved on since then.

 

[Horizon22]

Another twist. In this case could the stronger springs have pushed the problem onto less easily repaired/replaced components ?.

Of course, perish the thought, but maybe the permanent way needs to be kinder to the 80x ?. Are the culprits just a few rogue locations on the network ?. Not really the fault of the network of course and not a fault of the original spec it seems BUT maybe that is the cheaper solution overall ?.

Several test 80x trains with measuring equipment have been run acros the network over the past year, presumably to test this hypothesis. I don't think Hitachi were too happy at just accepting that the stock / build design is the only component.

 

[Peter Sarf]

Several test 80x trains with measuring equipment have been run acros the network over the past year, presumably to test this hypothesis. I don't think Hitachi were too happy at just accepting that the stock / build design is the only component.

Thanks for that info. Very true that Hitachi would be interested in the environment the 80Xs are subject to and a lot more interested than a mere bystander like myself !.

The new springs on the V of R coaches were also breaking and that was the problem I was dealing with. From memory the thicker section of the spring material and I think it was square section and not the original round section meant it had a higher stiffness and this resulted in higher stresses but I can’t remember the exact mechanism for failure. The point was that while on the face of it increasing the metal thickness should reduce the risk of breaking, not fully understanding the mechanism of failure meant that the solution used didn’t solve the problem.

The new springs were probably designed before 1970s and technology has moved on since then.

Sorry did not mean to suggest the stiffer springs were not THE problem. Just that, if they had survived, then the problem might have moved to another body part.

As you say, technology has changed a lot since the 1970s. Lots of different materials to consider these days - alloys etc. A bit of welding is not the same either - all those electronic components to upset and of course different welding techniques.

Gone are the days of Gaffer tape !.