Class 800 - Updates

[jim.snowdon]

The driver wont start sorting anything out without first contacting a few people and receiving the correct permissions.

 I agree with that but what happens when the computer says no?

 

Its all well and good sitting behind a keyboard stating this shouldnt happen and that shouldnt happen but what if it does happen, it has to be dealt with but if you have pulled all the levers and pushed all the buttons but the computer is still saying no, then what?

I may be sitting behind a keyboard these days, but I have done my part in specifying rolling stock, and dealing with their problems, and dealing with the design and acceptance of new stock, so I think I've probably earned the T-shirt over the last 40 years. Oh, and that includes driving trains, albeit test trains, as well.

 

The object, which should never be forgotten by everyone from train drivers to chief executives, is to keep the railway running and the passengers, or freight, moved from A to B. That is what they have paid for.

 

Jim

[royaloak]

I may be sitting behind a keyboard these days, but I have done my part in specifying rolling stock, and dealing with their problems, and dealing with the design and acceptance of new stock, so I think I've probably earned the T-shirt over the last 40 years. Oh, and that includes driving trains, albeit test trains, as well.

 

The object, which should never be forgotten by everyone from train drivers to chief executives, is to keep the railway running and the passengers, or freight, moved from A to B. That is what they have paid for.

 

Jim

So you are saying you have no/limited experience of dealing with computer based traction, is that correct.

 

Unfortunately with the 800s if the computer says no then its game over, you can isolate anything you want but nothing is going to happen unless the computer says so.

 

We had the same issues at SWT with the 444/450s when they were introduced and it was only after some things were made less 'precise' that the computer would say okay when things went wrong.

 

You cant simply isolate lots of things and get brake release these days!

 

I am sure everyone dealing with this incident was fully aware of the delays being caused and were trying everything to get the bloomin thing moving, there are those blaming the staff on the ground for the delays but even when Hitachi engineers turned up with their laptops the computer was still saying no, what more could the train crew have done?

Edited December 22, 2017 by royaloak

[jjb1970]

A lot of software controlled safety critical systems are designed precisely to stop operators finding ways to keep them going if a fault is present. For very good reason.

[jim.snowdon]

So you are saying you have no/limited experience of dealing with computer based traction, is that correct.

 

Unfortunately with the 800s if the computer says no then its game over, you can isolate anything you want but nothing is going to happen unless the computer says so.

 

We had the same issues at SWT with the 444/450s when they were introduced and it was only after some things were made less 'precise' that the computer would say okay when things went wrong.

 

You cant simply isolate lots of things and get brake release these days!

 

I am sure everyone dealing with this incident was fully aware of the delays being caused and were trying everything to get the bloomin thing moving, there are those blaming the staff on the ground for the delays but even when Hitachi engineers turned up with their laptops the computer was still saying no, what more could the train crew have done?

No, it is not. But I would have a serious argument with those designing computer systems that behave this way, and those who fail to specify the requirement for recoverability. Trains need to be designed to be recoverable, just as aircraft have to be designed not to fall out of the sky when the computer has a mental breakdown. It isn't the fault of the train crew if they can't deal with the problem because they haven't been provided with the means; the failure lies with the engineers who failed to provide the means in the first place and/or those who drafted the specification against which the train was procured.

 

I am, shall we say, anything but impressed with yesterday's performance.

 

Jim

Edited December 22, 2017 by jim.snowdon

[phil-b259]

 

The object, which should never be forgotten by everyone from train drivers to chief executives, is to keep the railway running and the passengers, or freight, moved from A to B. That is what they have paid for.

 

Really? Given the actions of the DfT over the past few years I was under the impression that getting passengers where they need to be was actually quite low down the list when it comes to 'objectives'. Making Politicians look good / ram through their ideologically driven 'initiatives' seems to be what matters in DfT land these days.

[The Stationmaster]

I presume the argument was that out of 12 units, 2 failed in service and 1 didn't even make it out of the depot.

 

So 9 out of 12 survived the day which is 75%.

 

Availability and Casualty rates (failures in traffic) are tow very different things.  Availability comes in two measures - the first is the percentage of the fleet which is actually available for traffic (which we don't know) and the other is the percentage delivered against diagram - that was 11 sets out of 12 so c.8.3% was lost meaning that fleet availability on the day was 91.7%.

 

Of the 11 sets in traffic 2, to our knowledge, were casualties and the fact that they became casualties is nothing to do with the availability figure but a lot to do with casualty/reliability measures and relevant statisytics (which is hopefully also defined in the contract between the train provider and DafT.

But, on any train, it is the driver who is one person who is available to start sorting out the problem and use whatever means are provided to allow it to be isolated and the train got moving again. But, there is a responsibility on the part of whoever is procuring the train to include the necessary requirements about recoverability in the contract specification, otherwise it won't happen.

 

What, as a train operator you shouldn't do, is to simply allow the railway to be stopped by your broken down train whilst you wait for the railway equivalent of the AA to turn up. The sad part is that in an environment where the commercial considerations take precedence, a train operator can take just that position if the cost of the delay penalties and reputational damage are considered to be tolerable.

 

Jim

 

I'm sorry Jim but we're talking about a train which is heavily software dependent and where the fault guide is more than likely part of that software.  All a Driver can do is guided by what the fault guide says and if things go beyond that it is then out of the Driver's hands. The obvious remedy on such a train is to have a software engineer on board if the fleet is suffering software reliability problems during the introductory phase (which is exactly how Eurostar managed a similar situation and which at times almost led to trains being 'driven' by an engineer using a laptop computer to sort out problems in the train's software.

 

No way do we want Drivers opening up and 'doing things' to what should be a 'sealed' and protected software system.  The problem with these trains is that the operator is not responsible for train provision and maintenance - that is done by a contractor.  At Eurostar things simpler in that it provided its own trains from its own depots and had its own engineers who knew the train's systems - GWR can't do that even if it  happened to have people with the necessary technical expertise because Hitachi are responsible for the trains.  (And maybey Hitachi, being a 'world class manufacturer' expected its trains to work out of the box and might well be surprised that for whatever reason there are software problems affecting them?  That we simply don't know but I bet there is egg on a number of faces and quite possibly a shortage of people to deal with any problems by riding on trains in traffic).

 

And let's not forget one of the basic things about this train - it was effectively procured by DafT and they are the people who should have made all the contract provisions for whatever circumstances were likely to arise - in association with the train operators.  As ever I'm sorry but if experienced engineers and operators aren't consulted during the specification and design stages then you won't get everything the everyday railway needs however if - as implied by one earlier post - an ordinary assisting loco isn't able to control the brake on these trains then there is something seriously amiss in the design. 

Edited December 22, 2017 by The Stationmaster

[royaloak]

No, it is not. But I would have a serious argument with those designing computer systems that behave this way, and those who fail to specify the requirement for recoverability. Trains need to be designed to be recoverable, just as aircraft have to be designed not to fall out of the sky when the computer has a mental breakdown. It isn't the fault of the train crew if they can't deal with the problem because they haven't been provided with the means; the failure lies with the engineers who failed to provide the means in the first place and/or those who drafted the specification against which the train was procured.

 

I am, shall we say, anything but impressed with yesterday's performance.

 

Jim

Nowadays trains are not designed by train engineers are they, it is all logarithms and lines of code because they know better than us old lot, even though they have never been near the front of a train or trackside, and a safety system to stop the train will do just that, what they havent thought about is clearing the line afterwards.

 

Today I have had to isolate 3 safety systems and do temporary block working, all in the space of 4 hours, dont you just love simulator days. 

 

Just look at the removal of the requirements for a yellow end because the headlights are set so far apart and are a certain brightness, never mind all the slight curves on the network which means the staff on the track wont see the bloody thing coming, but hey, it gives the designers more scope with their dynamic liveries doesnt it! When the lack of a yellow end is mentioned in a report (not that it ever will be) I would hope the person who decided it was a good idea can state their reasons in Court!

Edited December 22, 2017 by royaloak

[D854_Tiger]

But, on any train, it is the driver who is one person who is available to start sorting out the problem and use whatever means are provided to allow it to be isolated and the train got moving again. But, there is a responsibility on the part of whoever is procuring the train to include the necessary requirements about recoverability in the contract specification, otherwise it won't happen.

 

What, as a train operator you shouldn't do, is to simply allow the railway to be stopped by your broken down train whilst you wait for the railway equivalent of the AA to turn up. The sad part is that in an environment where the commercial considerations take precedence, a train operator can take just that position if the cost of the delay penalties and reputational damage are considered to be tolerable.

 

Jim

 

I wonder if they tried switching them off and then back on again.

[HillsideDepot]

Really? Given the actions of the DfT over the past few years I was under the impression that getting passengers where they need to be was actually quite low down the list when it comes to 'objectives'. Making Politicians look good / ram through their ideologically driven 'initiatives' seems to be what matters in DfT land these days.

And not just on the rail side of things.

 

There was a new Bus Services Act this year, which didn't get much support from local authorities when it was in development, nor as it went through the House. Sure we (largely me in the case of the LA I work for) completed the questionnaires, replied to the consultations and so on, but the Act is a lot of nothingness really.

 

Now that the Act is law DfT have sent someone to visit asking us why we're not making use of the provisions! He assured us that "everyone else is" but was silent when we asked for examples we could consider, yet offered a long list of other councils he had visited/was visiting on his "sales drive". 

 

It really is worrying when you get to actually speak to these people and realise that they really are as removed from reality as their output suggests!

Edited December 22, 2017 by HillsideDepot

[royaloak]

  At Eurostar things simpler in that it provided its own trains from its own depots and had its own engineers who knew the train's systems -

Even after the Hitachi staff turned up with their laptops they couldnt get the train to play, what chance a driver with a pen and piece of paper?

[jim.snowdon]

Availability and Casualty rates (failures in traffic) are tow very different things.  Availability comes in two measures - the first is the percentage of the fleet which is actually available for traffic (which we don't know) and the other is the percentage delivered against diagram - that was 11 sets out of 12 so c.8.3% was lost meaning that fleet availability on the day was 91.7%.

 

Of the 11 sets in traffic 2, to our knowledge, were casualties and the fact that they became casualties is nothing to do with the availability figure but a lot to do with casualty/reliability measures and relevant statisytics (which is hopefully also defined in the contract between the train provider and DafT.

 

I'm sorry Jim but we're talking about a train which is heavily software dependent and where the fault guide is more than likely part of that software.  All a Driver can do is guided by what the fault guide says and if things go beyond that it is then out of the Driver's hands. The obvious remedy on such a train is to have a software engineer on board if the fleet is suffering software reliability problems during the introductory phase (which is exactly how Eurostar managed a similar situation and which at times almost led to trains being 'driven' by an engineer using a laptop computer to sort out problems in the train's software.

 

No way do we want Drivers opening up and 'doing things' to what should be a 'sealed' and protected software system.  The problem with these trains is that the operator is not responsible for train provision and maintenance - that is done by a contractor.  At Eurostar things simpler in that it provided its own trains from its own depots and had its own engineers who knew the train's systems - GWR can't do that even if it  happened to have people with the necessary technical expertise because Hitachi are responsible for the trains.  (And maybey Hitachi, being a 'world class manufacturer' expected its trains to work out of the box and might well be surprised that for whatever reason there are software problems affecting them?  That we simply don't know but I bet there is egg on a number of faces and quite possibly a shortage of people to deal with any problems by riding on trains in traffic).

 

And let's not forget one of the basic things about this train - it was effectively procured by DafT and they are the people who should have made all the contract provisions for whatever circumstances were likely to arise - in association with the train operators.  As ever I'm sorry but if experienced engineers and operators aren't consulted during the specification and design stages then you won't get everything the everyday railway needs however if - as implied by one earlier post - an ordinary assisting loco isn't able to control the brake on these trains then there is something seriously amiss in the design. 

We have what we have, and nothing is really going to change that. The point is that what we have is deficient in the context of a working railway for want of sensible thinking and, as you state, consulting and listening to those with railway experience.

 

Jim

[APOLLO]

Bring back the Westerns & heavy MK 1 & 2's- fantastic get aways from a standing start.No computers required - (not much electrical anything either !!!).

 

 

 

Brit15

[JeffP]

Sadly, like in so many walks of life, (think: local councils, MP's etc), the things we put in place to serve us end up as our masters. Computers are the latest in the line.

 

Is it me, or has anyone else been treated to, "It won't let me...." when asking for something over the phone? Bookings, insurance, car parts, paying a bill, you name it, "computer says no!" is the new ideal..........

[The Stationmaster]

Bring back the Westerns & heavy MK 1 & 2's- fantastic get aways from a standing start.No computers required - (not much electrical anything either !!!).

 

 

 

Brit15

 

Well we've got the 800s so all being well the thousands might come next

[Zomboid]

800s continuing the illustrious history of Maybach engines on the western.

Though like D800/1/2 they're a little under done...

[rab]

Bring back the Westerns & heavy MK 1 & 2's- fantastic get aways from a standing start.No computers required - (not much electrical anything either !!!).

 

https://www.youtube.com/watch?v=FUyNP1-21Ic&feature=related

 

https://www.youtube.com/watch?v=5FDC0JFNqAI&feature=related

 

Brit15

Now that is what you call a train!

[HillsideDepot]

1A29 (1809 WSM - PAD) observed at Chippenham with the front 3 cars showing completely orange (every LED lit) side displays, while cars 4-10 all had a chequer board pattern of orange and black. If one unit had all orange and the other chequers that would perhaps be understandable, but mixed, but not random, seems very odd. Minor in the scheme of things, and not really affecting performance, but disappointing that they can't get it right. All the more so when the local family run bus company serving Chippenham has "MERRY CHRISTMAS" programmed as the "Not in Service" destination display just for this week.  

 

I also noted that the Down train was advertised as "front 8 coaches for Bath Spa" rather than 9 as last week. I assume works there have further reduced the available platform length.  

[jjb1970]

Systems are increasingly reliant on computer control systems. In many cases it is the only way to get the required performance, for example modern  engine power, efficiency and emissions performance couldn't be achieved without modern engine control systems. Safety critical systems are designed to try and engineer out the potential for human factors to result in a catastrophe. Automated operation can achieve big improvements in capacity and consistency in some processes. And a primary driver has been reducing manning levels in many industries. Either way, computer controlled systems won't be going away, quite the opposite as autonomous technologies mature. 

Not all "software faults" are software issues, often the software is doing its job and preventing operation with a physical fault or a hazardous situation.

[Coryton]

Not all "software faults" are software issues, often the software is doing its job and preventing operation with a physical fault or a hazardous situation.

 

Agreed, though in this case it seems that either the software failed far too safe or there were hardware faults on the train. I presume the software isn't designed to react to a problem with the overhead wires by refusing to run the diesel engines. 

 

My experience of Japanese engineering is of extreme conservatism - if they say something will do a job they are very very sure that it will, so I'm a little surprised at how things seem to be turning out here.

[RJS1977]

 

 

My experience of Japanese engineering is of extreme conservatism - if they say something will do a job they are very very sure that it will, 

 

Honda F1 engines, anyone?

[PaulRhB]

I wonder if they tried switching them off and then back on again.

That's been the standard fix with the 66's when they get the 'blue screen' took a while to spread but became a standard question when a 66 driver told you he'd failed.

On the positive side at least whatever fault caused it and then the reactionary issues can be addressed. Much better than finding out with the whole fleet in service and having to withdraw lots from service.

[The Stationmaster]

Agreed, though in this case it seems that either the software failed far too safe or there were hardware faults on the train. I presume the software isn't designed to react to a problem with the overhead wires by refusing to run the diesel engines. 

 

My experience of Japanese engineering is of extreme conservatism - if they say something will do a job they are very very sure that it will, so I'm a little surprised at how things seem to be turning out here.

 

I wonder (as Ron did) if part of the problem in this instance was down to the overhead?  The ADD seems from various things I've heard to be rather sensitive to several influences and far more sophisticated than simply being an 'overheight therefore I must lower the pan' piece of kit, and it also applies to raising the pan as well.  So - being wholly speculative - I suppose there is a possibility that it might be too finely tuned for various conditions encountered on the route?  However there are obviously from what observers in this thread are reporting what appear to be other TMS issues on these sets and as ever with such things they will no doubt take time to resolve - it took Class 373 trains c.4 years to get up to Vers.16 (yes, sixteen) of their software so getting things sorted on the Class 80X trains might take a while yet.

[Coryton]

I wonder (as Ron did) if part of the problem in this instance was down to the overhead?  The ADD seems from various things I've heard to be rather sensitive to several influences and far more sophisticated than simply being an 'overheight therefore I must lower the pan' piece of kit, and it also applies to raising the pan as well.  So - being wholly speculative - I suppose there is a possibility that it might be too finely tuned for various conditions encountered on the route?  However there are obviously from what observers in this thread are reporting what appear to be other TMS issues on these sets and as ever with such things they will no doubt take time to resolve - it took Class 373 trains c.4 years to get up to Vers.16 (yes, sixteen) of their software so getting things sorted on the Class 80X trains might take a while yet.

 

I would have thought the fact that the train seemed to have shut down completely following the ADD activation is more of a problem then the fact that it activated in the first place.

[jim.snowdon]

I would have thought the fact that the train seemed to have shut down completely following the ADD activation is more of a problem then the fact that it activated in the first place.

Exactly. The Train Management System not only seems to have gone into meltdown over what is really a trivial event, but did things it really ought never to do, like deciding the 10 car train was only 5 and rendering the complete train immovable. I appreciate what has been said about not wanting people (drivers) to interfere in safety critical systems, but sensible practice with safety systems is either to provide a level of redundancy so that a single failure is not disabling, or to provide bypass circuits protected by seals that, in emergency, can be used to get the train moving under restricted control, usually limited speed, so that it can be got out of the way and/or to a place where passengers can be disembarked. It is often practice to provide a limited number of hard-wired control circuits so that failure of the electronics to behave (as in this case, it would appear) some degree of control can still be exercised. Aircraft engineers would traditionally use a minimum triplication of control systems, with a two out of three voting system; the computer based interlockings used for signalling follow the same principle.

 

A question that perhaps ought to be asked is whether, if the TMS could reconfigure itself as a 5-car train, what happens in the unlikely event of coupler failure between the two units?

 

Jim

[jjb1970]

Certainly sounds like perhaps they should review the FMEA and fault tree analysis (assuming these studies were completed).