Entire 800/801/802 fleet stood down for safety checks

[An_Engineer]

Having worked as an engineer in aerospace + other applications with high stressed components, and also having done lab-based fatigue experiments, I think their is too much faith placed in the answers which FEA packages produce to ensure stresses which are "acceptable" and which are not assumed to have cyclic excursions. In aerospace, there will be ground based fatigue rigs of substantial parts of the aircraft which have accumulated a lot more stress cycles than the prototype aircraft being flown by test pilots; that way if anything untoward happens to the rigs on the ground, a warning can be flagged up in good time.

To chime in on this topic and to build on your excellent reply, I've done a fair amount of finite element (FE) modelling and simulation in my time. I can say with confidence that creating simulations and interpreting results is a dark art. You need to be incredibly careful with you inputs and assumptions (eg load applied, boundary conditions, contact conditions, material behaviour), and you also need to be very careful with interpreting your outputs (eg what type of stress value are you reading, are areas of high stress "true" values or a mathematical singularity, how do you interpret fatigue from stress results etc). The best way we found was to use FE as a tool to identify areas of concern during the design phase so that design mitigations/extra testing could be planned for in later physical testing. It is actually a really useful tool, but if you solely rely on FE results you're a fool.

For example, is it possible simulate an entire train using FE? No, absolutely no, not in any way. To accurately model something of that size and complexity would probably take several super computers a very long time, and the results (if it even managed to solve it) would probably be garbage as there would just be too many assumptions on part interactions. So to model our jacking points you would just take a small snippet around the jacking point, and make load assumptions about what the rest of the train is doing and how that load is transmitted to the area in question. Often this will lead to simplified simulations where only simple directional loads are applied, and then fatigue criteria are overlayed on the results (you can estimate fatigue life from that if you know an approximate loading amplitude and load frequency). It can eliminate gross miscalculations, but there is so much extra going on in real life that it is not a catch all solution.

People may be surprised by how much us engineers don't actually know, but we know we don't know them so we assess risks and plan testing to compensate. Inspection and maintenance regimes are part of this as well.

 

[O L Leigh]

How long does this knowledge last officially.

6 months after last driving a type of train or a route is the time which a drivers knowledge will officially lapse.

The issue becomes how you roster sufficient work within the link to ensure that competencies are maintained. But even that is an over-simplification, as rostering this within the link is one thing but delivering it in reality is quite another once you start to take into account leave, sickness, shift preference, etc, etc. It would quickly become unwieldy (much like this discussion) and undeliverable to the extent that if the traction reserve was going to be formed of the outgoing stock you would end up with the old trains never actually disappearing from the route. And for what? It's an incredibly expensive insurance policy for an eventuality that may never occur.

**EDIT**

I mean, I'm not quite sure how LSL/whoever are intending to provide an alternative service in the midst of a pandemic. Either they are able to provide crews who are route and traction competent or they're going to have to come up with some Covid-safe alternative using route conductors.

At the moment we're not allowed to be two in a cab. The only exceptions to this are training bubbles formed of permanently paired instructors and trainees, both of whom are subject to regular Covid testing and temperature monitoring. Any swapping of trainees or instructors is subject to testing and a 10 day isolation period.

 

[LOL The Irony]

Sorry to appear lazy, but (assuming it is known) can someone summarise what the problem(s) with these units is(are) please?

Cracks in the bogie bolsters and jacking points.

 

[pistonvalve]

Are these lifting lugs welded to the main chassis? How often would the body be lifted from the chassis? If welded, are the welds inspected?

 

[Qwerty133]

BR had a dual-sourcing policy after BREL was privatised, but that didn't necessarily work too well.
There were problems with one or other of the split fleets for Super Sprinters (155/156) and Networkers (465/466) and others.
HSTs were equipped with two types of engine and electrical equipment.
Every repeat 80x order also reduces the overall fleet costs and improved productivity at the maintenance depots.

I would have thought that there being issues with one (but not both types) shows the policy to be working. If neither fleet consistently experiences any significant issues it becomes something of a needless redundancy that increases cost and if both fleets experience issues it hasn't worked but in those cases the other fleet was able to remain in service whilst the problems were solved reducing the impact of an actual issue. Going further it could also be argued that backwards compatibility with existing stock can have a similar effect depending on the issues involved as happened when the 156/8s were hybridised to get around issues with the 158 fleet.

 

[anthony263]

The I newspaper says first group are delaying their Edinburgh open access operation now because of the faults

 

[AdamWW]

People may be surprised by how much us engineers don't actually know, but we know we don't know them so we assess risks and plan testing to compensate. Inspection and maintenance regimes are part of this as well.

Indeed. FEA is great for guiding design. But not to be relied on to show you got the design right.

At the moment we're not allowed to be two in a cab. The only exceptions to this are training bubbles formed of permanently paired instructors and trainees, both of whom are subject to regular Covid testing and temperature monitoring. Any swapping of trainees or instructors is subject to testing and a 10 day isolation period.

Given the currently very low rates of coronavirus and imminent relaxation of the rules on meeting people indoors, maybe it's time to re-think this?

 

[swt_passenger]

Are these lifting lugs welded to the main chassis? How often would the body be lifted from the chassis? If welded, are the welds inspected?

There is no chassis, it’s a monocoque construction. If they need to they lift the whole train, for a 9 car this would be 36 coordinated jacks.

 

[O L Leigh]

Given the currently very low rates of coronavirus and imminent relaxation of the rules on meeting people indoors, maybe it's time to re-think this?

And perhaps it will. But right now the restrictions still apply, and this is likely to be pertinent to how a replacement service is provided this week.

 

[edwin_m]

Hmmm. Are we then seeing a clear pattern where all trains above a certain mileage are too badly affected to run and those under it are OK?

I didn't think that was the case.

I don't think there's any clear picture yet for us outsiders and possibly not for insiders either. But for any design fault it's very likely that different examples of the same design subjected to the same conditions will fail in the same way at about the same time. That may not be the case if it's a manufacturing problem, but with the bodyshells concerned coming from several different sites I don't think that's likely.

 

[ainsworth74]

Are these lifting lugs welded to the main chassis? How often would the body be lifted from the chassis? If welded, are the welds inspected?

They're monocoque construction (integrated chassis and body) so I believe these are to life the vehicle off its bogies and to get underneath it more easily to maintain bits, take things off, etc etc.

 

[millemille]

Indeed. FEA is great for guiding design. But not to be relied on to show you got the design right.

Best you stop travelling on any trains in the UK then, because it's what's used to demonstrate compliance with the GM/RT 2100 standard...

 

[Nottingham59]

photos of EMRs 810s being welded by Hitachi Engineers, with lines like “Grab your Popcorn and Settle In

That EMR email is referring to a video, which shows the stir-welding process for the body panels, on YouTube:

 

[millemille]

That EMR email is referring to a video, which shows the stir-welding process for the body panels, on YouTube:

No it doesn't. It show MIG welding.

 

[AdamWW]

Best you stop travelling on any trains in the UK then, because it's what's used to demonstrate GMRT2100 compliance.....

Interesting.

I should have qualified my response by saying "in my experience", which (not in rail transport) has aligned with the view expressed above by An_Engineer that you can't rely on FEA alone, for the reasons they gave.

 

[Grumpy Git]

No it doesn't. It show MIG welding.

Agreed, FSW is an entirely different process altogether, see here for an explanation:

Edited with a better video, jump to 1m 30s for the actual process in action.

 

[fgwrich]

That EMR email is referring to a video, which shows the stir-welding process for the body panels, on YouTube:

Indeed it does, and that I’m aware of. But it is perhaps, unfortunate timing at best given the issues Hitachi are facing regarding problems with its own welding on its own Kasado built fleet.

 

[Stathern Jc]

Agreed, FSW is an entirely different process altogether, see here for an explanation:

That's really interesting, thanks.
(It's 40 years since I was an engineering student, and this hadn't been invented then )

 

[Wilts Wanderer]

It occurs to me that age and mileage of a trainset isn’t exactly the same thing. Everyone seems to be talking about cracking relating to how long a train has been in service, potential implications of rough riding or poor track or whatever. However the cracks are located on/around the jacking points. Could the severity of the fault somehow be related to how many times a particular vehicle has been lifted?

It reminds me of the fatigue cracks that ultimately crashed a Douglas DC10 in the 1970s, where the airline had been using a forklift to remove the engines + pylon as a complete assembly rather than doing it separately with specialised lifting gear per the manufacturer’s guidelines. Not suggesting for a moment that HRE haven’t followed correct procedures but I wonder if the lifting process is somehow a contributing factor.

 

[Starmill]

Whilst Pewsey railway station will not be served by any train services, road transport will operate between Pewsey and Newbury.
.

At least people at Pewsey aren't totally stuck then.

 

[millemille]

It occurs to me that age and mileage of a trainset isn’t exactly the same thing. Everyone seems to be talking about cracking relating to how long a train has been in service, potential implications of rough riding or poor track or whatever. However the cracks are located on/around the jacking points. Could the severity of the fault somehow be related to how many times a particular vehicle has been lifted?

It reminds me of the fatigue cracks that ultimately crashed a Douglas DC10 in the 1970s, where the airline had been using a forklift to remove the engines + pylon as a complete assembly rather than doing it separately with specialised lifting gear per the manufacturer’s guidelines. Not suggesting for a moment that HRE haven’t followed correct procedures but I wonder if the lifting process is somehow a contributing factor.

Not "everyone" has been talking about mileage, you'll note I've described it as duty cycle. This includes everything that imposes "duty" upon the failed part.

I'm sure that one of the very first things looked into by Hitachi will be exactly what you've identified, a relatively simple but probably time consuming check to make going through the vehicle maintenance history to look for activities being completed that would have required lifting; power pack changes, bogie changes, wheelset changes, compressor changes, overhauls etc.

But this will also need to be done for vehicles which don't have any cracks as it is not significant if there are vehicles which have been lifted just often but without showing cracks.

 

[43066]

I was sticking the cabs of 465's on with Sikaflex in 1991, so ner!

If I was in any way responsible for the fit and finish of networker cabs I wouldn’t go around admitting to it .

Non-provision of the service paid for is unlikely to constitute discharge of the TOCs responsibilities.

NRCOT n.28.3 would in my view cover most passengers affected by this mess: there is an express contractual right to rebook for a later date in the event of delays.

If this were to reach a court, I would expect delay to be interpreted to include outright cancellation.

If a TOC were to be sued for non-compliance, I would advise them to settle at whatever price to avoid it reaching a formal judgment.

Fair enough (and with apologies for last night’s intemperate postings). Given the amounts at stake, I’m genuinely surprised there is much of a market for legal advice to individuals pursuing TOCs for losses stemming from cancelled trains.

Presumably for some it becomes a point of principle?

 

[bb21]

Best you stop travelling on any trains in the UK then, because it's what's used to demonstrate GMRT2100 compliance.....

I know it's a quick google away but you might wanna quickly explain what the standards are in your professional opinion for those less knowledgeable.

 

[millemille]

If I was in any way responsible for the fit and finish of networker cabs I wouldn’t go around admitting to it .

Ooooooooh! Saucer of milk for @43066......

 

[millemille]

I know it's a quick google away but you might wanna quickly explain what the standards are in your professional opinion for those less knowledgeable.

Ok, there are National Technical Rules for the UK railway undertaking as mandated by law in the shape of Railways (Interoperability) Regulations 2011. In this particular instance Railway Group Standard GMRT2100,"Rail Vehicle Structures and Passive Safety" covers, as the name implies, the requirements for the structure of the vehicle. It's issue 6 now, but was issue 5 when the class 80*'s were subject to scrutiny.

GMRT2100 in turn refers to a number of British and European standards for the calculation methods, load cases, duty cycles, material data sources, welding data sources etc. and is also supported by a Rail Industry Standard - which didn't exist at the time of 80*'s - RIS-2780 "Rail Vehicle Structures" which gives guidance on areas not covered by the national technical rules.

Every new train, or modification to an existing train, has to demonstrate compliance with the applicable standards in force at either the time of submission for review or when the order was placed (in the case of new trains or other large, long term capital projects).

In simple crude terms, the designers of the trains don't have to decide how "strong" the train needs to be as the standard tells them. They have to demonstrate, in this case through using computer aided modelling known as Finite Element Analysis (FEA), that their design complies with the required "strength".

I design, in conjunction with my design engineers, bits of trains that have to demonstrate compliance with GMRT2100. I do long hand calculations the old fashioned way to help with the design process and then the CAD package has FEA functionality but once the design is complete it goes to a specialist structural analysis consultancy who use a full monty FEA package to do the full, formal analysis to demonstrate compliance with GMRT2100.

This is the established, accepted methodology for at least the last 20 years in the UK. The FEA packages are not rail specific and panels of clever people, far far cleverer than the likes of me, have written the standards for others to follow. The UK rail industry demonstrates safety through compliance with standards.

​

​

 

[1D53]

800014 now also been stopped, now 800006 & 800009 in traffic.

 

[stuartl]

It occurs to me that age and mileage of a trainset isn’t exactly the same thing. Everyone seems to be talking about cracking relating to how long a train has been in service, potential implications of rough riding or poor track or whatever. However the cracks are located on/around the jacking points. Could the severity of the fault somehow be related to how many times a particular vehicle has been lifted?

It reminds me of the fatigue cracks that ultimately crashed a Douglas DC10 in the 1970s, where the airline had been using a forklift to remove the engines + pylon as a complete assembly rather than doing it separately with specialised lifting gear per the manufacturer’s guidelines. Not suggesting for a moment that HRE haven’t followed correct procedures but I wonder if the lifting process is somehow a contributing factor.

The part about the lifting process was my first thought, if all the jacks don't lift evenly then I would think it would add extra stress onto the body shell. May well turn out to be a combination of several things, none of which in themselves would cause a problem.

 

[Failed Unit]

Have LNER had another hit of failed units. More cancellations tomorrow including the entire Lincoln service.

 

[43066]

Ok, there are National Technical Rules for the UK railway undertaking as mandated by law in the shape of Railways (Interoperability) Regulations 2011. In this particular instance Railway Group Standard GMRT2100,"Rail Vehicle Structures and Passive Safety" covers, as the name implies, the requirements for the structure of the vehicle. It's issue 6 now, but was issue 5 when the class 80*'s were subject to scrutiny.

GMRT2100 in turn refers to a number of British and European standards for the calculation methods, load cases, duty cycles, material data sources, welding data sources etc. and is also supported by a Rail Industry Standard - which didn't exist at the time of 80*'s - RIS-2780 "Rail Vehicle Structures" which gives guidance on areas not covered by the national technical rules.

Every new train, or modification to an existing train, has to demonstrate compliance with the applicable standards in force at either the time of submission for review or when the order was placed (in the case of new trains or other large, long term capital projects).

In simple crude terms, the designers of the trains don't have to decide how "strong" the train needs to be as the standard tells them. They have to demonstrate, in this case through using computer aided modelling known as Finite Element Analysis (FEA), that their design complies with the required "strength".

I design, in conjunction with my design engineers, bits of trains that have to demonstrate compliance with GMRT2100. I do long hand calculations the old fashioned way to help with the design process and then the CAD package has FEA functionality but once the design is complete it goes to a specialist structural analysis consultancy who use a full monty FEA package to do the full, formal analysis to demonstrate compliance with GMRT2100.

This is the established, accepted methodology for at least the last 20 years in the UK. The FEA packages are not rail specific and panels of clever people, far far cleverer than the likes of me, have written the standards for others to follow. The UK rail industry demonstrates safety through compliance with standards.

​

​

Cracking, informative content on here, thanks.

I’ll even forgive you the networkers.

 

[73128]

Can they not be inspected by a "man in a van" in the yard at Worcester, and only go to Bristol if repairs are needed?

There's one abandoned in the up sidings at Oxford as well.