Migrating The Interface

Discussion relating to the operations of real railways together with the experiences of the people who work (or have worked) on them.

Moderator: Moderators

Locked
User avatar
CSRZiyang
Getting the hang of things now
Posts: 24
Joined: Wed Sep 13, 2006 7:50 pm
Location: Bishop's Stortford
Contact:

Migrating The Interface

Post by CSRZiyang »

I'd like to use this forum to bounce an idea around.

Currently, the engineering interface between the infrastructure provider (Network Rail) and the train operator is at two particularly tricky places, the wheel / rail and the platform / train.

What if this interface were to be migrated upwards, so that Network Rail were responsible for these interfaces (effectively taking responsibility for the "nuts and bolts" of trains) and train operators simply held influence over the design of the exterior livery and the train interior?

Over the next ten to fifteen years, there will need to be two large fleet replacement projects; IEP (HST2, ICE, or whatever) and a large DMU programme, to replace life-expired Pacers / Sprinters.

If designs were developed and prototypes manufactured, with Network Rail (or maybe a not-for-profit "Network Trains" organisation) owning the intellectual property, then tenders for series build could be invited from manufacturers.

The benefits of this approach are potentially numerous: -
  • There's a good chance that the resultant designs would be track-friendly, as Network Rail would be incentivised in a way that no train builder could be unilaterally;

    There's a good chance that track-friendliness would also bring low weight, thus providing good energy efficiency;

    The simpler interfaces would remove engineering risk, meaning that (in theory) leasing companies should be able to offer lower lease rates;

    Train operators could concentrate their minds on the softer parts of the train; i.e. passenger comfort, cleanability and marketing issues like branding, the catering offer etc;

    The train builders would be competing on a level playing field, all offering prices against the same design, making tender evaluation easier;

    A large build could be split between several manufacturers, without any risk of the BREL / GEC Alsthom Networker situation (complete lack of interchangeability across the two parts of the 465 "fleet").
I'd be interested to hear the views of forum members.
mattvince
Very Active Forum Member
Posts: 1739
Joined: Sun Apr 27, 2003 8:48 pm

Post by mattvince »

Make that three fleet replacements - the large fleets of 1970s/1980s suburban EMUs, starting with Classes 313 and 315. And what of Thameslink? Clearly, if the Thameslink scheme is to receive new stock, then it makes sense to use the same type (or variation thereof) to replace the 313s on the Moorgate-Welwyn/Hertford lines. By 2030, and assuming a 35-year lifespan, there is a need to replace about 1500 DMU vehicles and 3500 suburban EMU vehicles. The pace of replacement could well be higher if units cannot be derogated from or modified to meet RVAR standards by 2020.

The concept sounds fair and reasonable, but I can see there being conflicts between Network Rail and the DfT, who also seem to be intent on building the train-set as well as playing with it (HST2/IEP has been rumbling on for the last three or four years - VXC had Voyagers in service within four years of ordering). There could also be competition issues - would such a rigid specification fall foul of various domestic and European regulations over procurement? Would there be a 'compulsion' to acquire the NR-spec units, over an operator-specified design (if the operator was to favour a non-standard design, or if the operator had reservations over one of the chosen suppliers)? What of the risk of stagnating development - would we be solely limited to evolutions of a few standard designs (would a 'tweaked' version of a 2015 design still be suitable in 2050?), and what if the rail industry becomes aware that it needs to radically change its designs (to adopt new power sources, or double-deckers, for example?). I have nothing against evolving designs - just as long as there is no repeat of the four-tonne increase in unit weight between Classes 317 and 322!

There is a lot which could be done to standardise parts, principally by the full adoption of IEC 61375 (Train Communication Network) and extend its principles into other areas, so that a X's traction package can be plugged into Y's Train Management System, all in a bodyshell built by Z. That would allow operators to "pick'n'mix" the parts they want, and know the cost of the whole system. Many other changes could be done through Railway Group Standards, incentivised where necessary by the Rail Regulator.

Looking at the future direction, it might be that under a future Conservative government, the whole nature of the rail industry could be turned on its head, thus it could be that large, vertically-integrated regional operators, with either very long 'franchises', or true Railway Companies (in the traditional sense) would have both the ability to procure units without resorting to leasing companies, and would have the direct cost incentive to buy track-friendly, efficient designs. They could take this opportunity to specify and design the units which they, as owner-operators, need rather than the decisions of a centralised Design & Procurement Committee - and someone of your calibre should be well aware of what happens when things are designed by committees... :)
User avatar
Elojikal
Very Active Forum Member
Posts: 1135
Joined: Mon Mar 17, 2003 5:56 pm

Post by Elojikal »

Like the whole Networker concept this idea sounds fine in principle but would it work in practice?

What if a Network Rail standard was a bad one? Would a bad standard be forced upon all the train operators? As an example what if a standardised train management software solution based on Windows operating systems and with middleware based on proprietary technology and the actual train management software being poorly written and having many flaws such as memory leaks and poor error handling? What if the terms of the contract between Network Rail and the supplier meant that the manufacturers had limited or non-existent access to the code of the train management system and therefore were tied to the development cycle of the supplier. What if the supplier went out of business?

This would be fine is Network Rail were the architects of the train management systems and specified that they ran on strong and reliable operating systems and hardware (ie. not desktop operating systems running on 15 year old computing technology) and were also responsible for the development of the TMS themselves and thus directly responsible to the operators. Sadly as the IT decisions of many companies are heavily based on the marketing bullsh*t they receive from large commercial suppliers (who in the case of the public sector often cosy up to politicians and lobby them directly for lucrative software contracts) I could not see this being the case.

Secondly, what if Network Rail were to force homogeneous designs upon the operators? Part of the biggest problem with the Class 375 and 377 fleets of Electrostars is that they are a homogeneous design - jack of all trades and master of none. Although all modern multiple unit designs come from a base family of trains (eg. Coradia, Electrostar or Desiro) specialised adaptations of train designs are required for different services (long distance limited stop, stopping, inner suburban, outer suburban.) South West Trains have been able to procure a fleet of trains that meets their differing needs with the 444s, 450s and 458s. What if they were to have a single design forced upon them that had carriages of fixed length, fixed door formations, fixed corridor connection dimensions and to reduce wear on the track and the strain on the network power supply a fixed number of engines per unit giving a uniform amount of acceleration and braking?

Ideally this would not be the case but a centralised authority detached from the commercial concerns of the operators yet not too detached from the political and financial concerns of the government could quite conceivably do so.

Thirdly, when you say track friendly designs, do you see the potential for a centralised authority to impose a fleet of multiple units on an operator when a fleet of locomotive based designs might be more appropriate for their services?

Fourth, could a single carriage design conceivably be forced upon all operators regardless of operating conditions? Trains with carriages whose designs are based upon or similar to the MK3 body profile are often spacious and comfortable inside (Desiros, for example.) Those with carriages that more closely resemble the MK4 design even though they don't have tilting technology and most likely never will are often cramped and uncomfortable on the inside (Class 159s and Electrostars immediately come to mind.)

Fifth, would standardised designs insist upon a single supplier for train components? One of the biggest problems with the reliability of modern rolling stock is that the companies that build the trains a) often contract out much of the work to third party suppliers and b) by extension not responsible for many of the parts that go into the trains that they manufacture.

Sixth, I had some other point but I've been waffling on so much I forgot what it was... Never mind I'm sure I'll remember!
User avatar
johncard
Very Active Forum Member
Posts: 1285
Joined: Mon Apr 05, 2004 9:42 am
Location: Sheffield

Post by johncard »

I tried posting this last night, so it might already be covered. Please bear the below in mind that I have not long finished A levels.

Sounds like a good idea, but may give rise to the question - how far could train operators provide a better "product" to their competitors solely through internal furnishings, and how would this affect profits? Just as an example, would you rather travel from Doncaster to York on a crowded Voyager with under floor engines but modern interior, or a quieter, half empty Mk 3 with aging decor?

Of course an extension of this could be vertical integration, but then there are all sorts of issues surrounding that as well - which companies would own what routes in respect of freight? In the “big four” days it was the same companies that provided for passengers and freight, but now the FOCs (presuming they still existed) would have to run over railway companies tracks – would these companies have the right to operate freight services of their own? Would track access charges be regulated? Would scheduling be regulated? Would freight-only routes be owned by an FOC, Network Rail (or successor), or the company/ies owning the connecting lines? I understand the Tories have been seriously considering this.

John
User avatar
thenudehamster
Very Active Forum Member
Posts: 5029
Joined: Wed Sep 10, 2003 7:56 pm
Location: Somewhere in cyberspace
Contact:

Post by thenudehamster »

I think that we're in danger of delving too deeply into a problem that may not even be there if we take Elojikal's premise to a conclusion. Perhaps the general principles standard would be a good idea, leaving the detail design to the manufacturers.

As a small analogy, look at mains electricity in your home. There is a standard set of design parameters for certain critical aspects of three-pin plugs and sockets, switches, cables and circuit breakers, but the detail design of those items is up to the manufacturer. We therefore have fully moulded plugs and screwed together plugs, of different shapes, materials and colours, some with extra handles for the disabled, some not; cables and wires are attached in several dfferent ways; we have sockets with and without switches, rocker switches, toggle switches and dimmers - but any plug will fit any socket, and all switches work all circuits. Apply that sort of standard to railway vehicle design and you're in with a fighting chance - provided it's left to someone like the BSI to set the standard, not a politically controlled and motivated government agency....

No reason why you can't have certain 'standard' wheelset designs, bogie designs and characteristics, and basic seating and gangway standards - but that doesn't mean that all carriages have to be the same size and shape or that each one has to have 143 seats made of moulded plastic with a two inch foam cushion and pink upholstery, though they would have to have, for instance, a minimum standard size of seat squab and back height, rake, and 'squashability'.

It's a proposal worth considering.


BarryH - thenudehamster
User avatar
CSRZiyang
Getting the hang of things now
Posts: 24
Joined: Wed Sep 13, 2006 7:50 pm
Location: Bishop's Stortford
Contact:

Post by CSRZiyang »

Wow, I'm overwhelmed with the intelligence and foresight that exists in this place! I find it bizarre that a forum which majors on train simulation has such a wealth of real railway knowledge buried so deep down within it.

Many thanks for your responses. I am particularly interested in the vertical integration thoughts. Of course in the days of the "big four", inter-operability issues were dealt with by the Railway Clearing House, an organisation which began life as a financial settlement organisation but ended up as a major specifier and owner of a considerable amount of IPR, to anything from buffers to wheels to electrical connectors (still fitted to some modern vehicles!). In the current regime, there is no reason why RSSB shouldn't assume a greater role and become a detailed specifier.

On the subject of timing, let's say it takes two years for an election to take place and (assuming the Tories win), another three years for them to find their feet and come up with policies, another three years to untangle and reform the railways, and another four years (using the Voyager benchmark) to come up with stock, that puts us in 2018.... can we wait that long for the plethora of new build stock that is required? FGW certainly won't wait that long to replace their HSTs and will probably go their own way long before that.

On the matter of evolution and developing technology, the specifications need to accommodate this as far as is possible without a crystal ball. For example, a power car that is developed at day one as a diesel power car needs to have the body structure and roof profile designed so that it can accommodate a pantograph at a later date.

Everything needs to be as modular as possible, so that (for example), a diesel underfloor power pack (containing alternator and control equipment in the case of DEMU and cardan shaft output to bogie in the case of DHMU) can be simply removed and replaced with another widget that does the same job.

If the (say) DMU is an inner suburban unit, which does not have air conditioning (contentious, I know but we want to save energy, weight and increase reliability, don't we?) then it needs to have the roof pockets and ductwork in place, so that at a later date air conditioning could be fitted if the unit were re-deployed on longer routes or the climate changed more dramatically than anticipated.

Regarding body profile, I think we need to either take a step back and base everything on the 'go (almost) anywhere' Mk. III profile or make a push to enhance the structure gauge and give ourselves the luxury of double-deck capability. But, we do need to make a decision one way or the other.

On speed, I agree with Cliff Perry that 125 mph is fast enough. I also think that non-tilt at 125 mph is reasonable, even on the WCML. There is absolutely no difference between a driver slowing down to traverse a structure (either under a TSR or a PSR) and slowing down to take a bend that has been flattened during the WCML upgrade. Modern traction has much better acceleration and braking capaibilities than in the past and this should be used to the full. With GPS and train control technology advancing fast, the speed could even be automatically governed in the future.

On TMS, I think we are in danger of obsessing about this. I worked in the oil and gas industry for a number of years as a control and instrumentation engineer prior to joing the railway industry. In that industry, SCADA (supervisory control and data acquisition) and DAS (data acquisition system) have been commonplace for over thirty years. BUT, these systems are only used to collect data and carry out no front-line role, simply being used to advise personnel regarding system status. A seperate (usually hard wired) ESD (emergency shut-down) system has the task of shutting things down in a controlled and safe way in the case of an abnormal condition that puts either equipment or life at risk.

There is no reason why this principle shouldn't extend to trains and to a certain extent it does already. The problem is that we expect the TMS to do too much.

Traincrew should have TMS information on a 'need to know' basis. If the ESD system decides to shut down the system, then the driver needs enough information to be able to discuss with either a technical riding inspector or a remote technical support engineer (by phone) what the problem is, and what he or she can do to (in the worst case) either get the train moving and clear the line or call for traction assistance.

Over and above that, the TMS should sit in the background, quietly gathering data, which, if used properly should be able to give maintenance staff a great insight into how the train (and with a larger overview the fleet) is performing and how maiintenance processes might be tweaked to get more reliability.

Anyway, I've rambled enough. I look forward to continuing this interesting debate.
User avatar
Elojikal
Very Active Forum Member
Posts: 1135
Joined: Mon Mar 17, 2003 5:56 pm

Post by Elojikal »

thenudehamster wrote:I think that we're in danger of delving too deeply into a problem that may not even be there if we take Elojikal's premise to a conclusion. Perhaps the general principles standard would be a good idea, leaving the detail design to the manufacturers.
Firstly I get the impression that you thought I was referring to detailed design of trains such as the interiors, which I wasn't. ;)

The thing is, even with a general principles standard it could still end up being a bad standard. Maybe not even a bad standard but one which leaves a lot to be desired. In the early 50s when British Rail were looking to specify a standard coach design that would become the MK1 coach they settled on the Bulleid designed coaches used in the Southern Region's suburban 4-SUB and latterly EPB trains. They did this largely as a matter of convenience and expediency. The problem with this is that the SUB and EPB coach designs were based on a design from the mid 30s that had been intended for a class of trains whose production had been interrupted by the war. These designs themselves owed much to the previous coach designs of LBSCR stock dating back to the turn of the 20th century and as Southern often recycled rolling stock the newer stock could not be too radical in terms of design. By the time British Rail came to settle on the MK1 standard coach design it was a design that was essentially twenty years out of date yet came to dominant subsequent coach designs for the next thirty years. However around the same time Bulleid was designing coaches with seperate bodies and underframes and "slam doors" the LMS were already bringing into production suburban trains of an integral design and with sliding doors. If British Rail had based their standardised coach design on a more modern design such as this then the name "Clapham" might not be brought up so often in conversation when discussing the end of the MK1 era.

Of course in the future such a bad decision as this would most likely not be made again however I think it is a valid point about how a central authority can go wrong due to compromise when coming up with general design principles.

And British Rail came up with many bad bogie designs, some absolute shockers in fact. The design of bogies that are track friendly and give good riding quality and comfort seems to be more of an art than a science!

My arguments are not against a central body setting standardised design principles, I think we need them. My concern is with how the central body operates and whether or not inferior standards might be forced to prevail over superior alternative designs.


CSRZiyang wrote:On TMS, I think we are in danger of obsessing about this. I worked in the oil and gas industry for a number of years as a control and instrumentation engineer prior to joing the railway industry. In that industry, SCADA (supervisory control and data acquisition) and DAS (data acquisition system) have been commonplace for over thirty years. BUT, these systems are only used to collect data and carry out no front-line role, simply being used to advise personnel regarding system status. A seperate (usually hard wired) ESD (emergency shut-down) system has the task of shutting things down in a controlled and safe way in the case of an abnormal condition that puts either equipment or life at risk.

There is no reason why this principle shouldn't extend to trains and to a certain extent it does already. The problem is that we expect the TMS to do too much.

Traincrew should have TMS information on a 'need to know' basis. If the ESD system decides to shut down the system, then the driver needs enough information to be able to discuss with either a technical riding inspector or a remote technical support engineer (by phone) what the problem is, and what he or she can do to (in the worst case) either get the train moving and clear the line or call for traction assistance.

Over and above that, the TMS should sit in the background, quietly gathering data, which, if used properly should be able to give maintenance staff a great insight into how the train (and with a larger overview the fleet) is performing and how maiintenance processes might be tweaked to get more reliability.
My main concern about TMS isn't really a "ground level" issue so much as a network issue. We have a situation in this country where Virgin who have realised the limitations of their Voyager fleet can't augment this fleet with a number of Meridian's that Midland Mainline trains suddenly found they have no use for because the two fleets of trains used different and incompatible TMS. The same problem exists with the different fleets of Electrostars where different versions of TMS exist. What we really need is a standard train management system that is used on all trains in this country. We need a single centralised authrority to make this happen but I fear that if that authority wasn't at arms length from the government then it might approach the design of a standardised TMS from a technocratic viewpoint.

And I agree about the traincrew. A train driver only needs to know how to operate a train. He doesn't so much need to know how it operates. He certainly needs an understanding of the mechanical side but he doesn't need to concern himself with why the TMS is being uncooperative and not opening the doors.

However you go on to say that we are quite rapidly approaching a future where ATO systems are scaled up to fully take over the operation of main line train services. If this were to be the case then if an operator remained on the train even in a safety role then he would need to have a greater knowledge of the works of the TMS.

For the record I don't think GPS is the appropriate technology for the primary automatic operation of trains. It isn't reliable enough in urban areas to be failsafe. Rather I see it as a component of an advanced ATO of the future being utilised in a two way form whereby the TMS sends its position, the status of the train and the operating conditions back to a central computer. This central computer which would have the capacity to compute hundreds of alternative pathings and timings (much like computers designed to play chess do) and make adjustments to the service of the train in operation based on problems that lie ahead such as late running trains, failed points, lights and so on. This network information would then be sent back to the train's TMS and it would then respond to it by taking the appropriate actions.

PS. I don't find your suggestion of not having air-conditioning as contentious as I am someone who prefers fresh air to recycled air - especially if the train operator has a policy of not keeping its carriages cool in the summer months in order to reduce energy consumption! ;)
User avatar
CSRZiyang
Getting the hang of things now
Posts: 24
Joined: Wed Sep 13, 2006 7:50 pm
Location: Bishop's Stortford
Contact:

Post by CSRZiyang »

Elojikal wrote: My main concern about TMS isn't really a "ground level" issue so much as a network issue. We have a situation in this country where Virgin who have realised the limitations of their Voyager fleet can't augment this fleet with a number of Meridian's that Midland Mainline trains suddenly found they have no use for because the two fleets of trains used different and incompatible TMS. The same problem exists with the different fleets of Electrostars where different versions of TMS exist. What we really need is a standard train management system that is used on all trains in this country. We need a single centralised authrority to make this happen but I fear that if that authority wasn't at arms length from the government then it might approach the design of a standardised TMS from a technocratic viewpoint.
This is more of a commercial and project management issue than a technical one (although when purchasing follow-on build vehicles it would be helpful to specify that they should be fully compatible with those that preceded them!). The manufacturer should be brought to book and "encouraged" to sort out this problem. Whilst I represent a manufacturer, I still believe that "customer is king" and if the market were truly competitive, this would not be an issue and the train builder would sort the problem.
Elojikal wrote: However you go on to say that we are quite rapidly approaching a future where ATO systems are scaled up to fully take over the operation of main line train services. If this were to be the case then if an operator remained on the train even in a safety role then he would need to have a greater knowledge of the works of the TMS.

For the record I don't think GPS is the appropriate technology for the primary automatic operation of trains. It isn't reliable enough in urban areas to be failsafe. Rather I see it as a component of an advanced ATO of the future being utilised in a two way form whereby the TMS sends its position, the status of the train and the operating conditions back to a central computer. This central computer which would have the capacity to compute hundreds of alternative pathings and timings (much like computers designed to play chess do) and make adjustments to the service of the train in operation based on problems that lie ahead such as late running trains, failed points, lights and so on. This network information would then be sent back to the train's TMS and it would then respond to it by taking the appropriate actions.
Actually, I wasn't advocating full automatic operation at all, simply automatic speed governing when approaching curves with permanent speed restrictions for that type of stock (e.g. non-tilting trains on WCML curves). This would allow the driver to fully control the train, but would not allow him to exceed the safe maximum speed at given points. If GPS were not considered dependable enough on its own, it could be supported by distance measurement from a datum, using an on-board odometer, with the GPS providing regular inputs to cross-reference and auto-reset at (say) Euston and Manchester (or any given station stopping point).
Locked

Return to “Real Railway Discussion”