The Argument for an Enlarged Loading Gauge

[HSTEd]

As all the long timers in this forum will know, we regularly get threads regarding double decker trains - especially when some politician or other touts them as a solution to capacity problems.

The following arguments are often cited against conversion to GC gauge or GB+ to utilise standard european double deck stock:

• The loading gauge is wider below the platform level and thus would require massive platform works, and likely dedicated platforms
• The low floor on the lower deck would prevent gangways being provided on both decks, and would prevent a flat lower deck, and would thus inhibit disabled access and loading operations
• The vast majority of the cross section of the loading gauge would be filled with passenger accomodation or with equipment like bogies, rendering it difficult to find space for traction gear without sacrificing train length
• In addition to platforms, large amounts of equipment adjacent to the track might have to be moved, and massive painstaking reclearing works would be required

One thing that I have not often heard suggested, is that we could develop a specialist loading gauge that would be identical to the British loading gauge as far high as the cantrail, and would then simply continue upwards for a much larger distance, likely terminating with a flat roof more compatible with modern rolling stock design practices.

This would have significant advantages over a conversion to a continental style loading gauge:

• The loading gauge would be identical to traditional British practice below platform level, and can thus use the same platform structures as conventional platforms
• Conventional gauge self powered trains would be able to use large gauge lines with no modifications, and electric trains would only require an extended travel pantograph
• Once you commit to a new custom loading gauge, there is little reason to restrict the vertical clearance to the ~4700mm of the GC gauge, which in any case would not be sufficient to enable a second deck above the normal height one. In any case selecting the largest possible height compatible with stability would enable expanded freight operations - 6150mm
• As stated above, a clearance of 6150mm would enable expanded freight operations - for example double deck container operations on cleared lines. Drastically increasing the number of containers that can be hauled in a path - an important consideration on the British railway, and potentially drastically reducing costs by increasing attainable train weight.
• Virtually all lorries commonly used [Motorways are only built for ~5m vehicles in any case] on public roads could be carried on standard flat wagons, permitting rolling highways and piggyback trailer operations to be conducted far more easily and cheaply than elsewhere. This would enable railways to help reduce HGV journeys more effectively than conventional intermodal operations.
• As there would be no difference from traditional loading gauge below the cantrail, far fewer modifications would be required, especially outside of large cities where there are only relatively few overbridges and tunnels.
• The maximum cant permitted on the railway is only roughly 150mm, which translates to a tilt of six degrees, which translates to a track overhang of 610mm. Roughly 300mm more than in traditional lines. This is only a single foot and it is likely that most curved track sections would easily meet this specification without modification.
• Trains would have level lower decks, and would be gangwayed (inside the unit) on both decks, this would permit far greater lateral movement through the train and would remove many of the disadvantages seen in continental double deck trains. For example only one toilet (the disabled one) would be provided on the lower deck of a train regardless of length, all others would be on the upper deck, along with lower passenger-density areas such as catering bases, first class and similar.
• Full width cabs could be provided on the upper deck, permitting gangwaying between units on the lower deck without sacrificing driver comfort and visibility
• As the lower deck would be at the same level as traditional trains, double deck units could operate in multiple with existing classes of single deck units as required. And traction gear could be positioned below the floor the train, as on exisitng rolling stock

So in other words - although the vertical clearance is extreme there are actually relatively few bridges per mile on the network as a whole, and the only real problem is that electrification equipment would have to be rebuilt.
And since the ECML is going to have lots of structures replaced anyway......

But it's just a thought.

 

[Bald Rick]

Couple of points.

Most Intermodal trains are now limited by trailing load as much as length. Double stacking certainly won't mean double the boxes per path.

Also, I'm not sure how could fit 2 X ISO 9'6" containers in a loading gauge of just over 6metres, especially if OLE is involved.

Finally, all those new over bridges that have been built over the last 50 years to GB+ would need rebuilding, including every single Motorway bridge. As well as all the other bridges and tunnels of course, and all the OLE (almost all of which would need new masts). Rebuilding tunnels is not easy, as Farnworth, Ipswich and Southampton show us. And the latter two were being given an extra 300mm, not nearly 10 times as much.

 

[HSTEd]

Couple of points.

Most Intermodal trains are now limited by trailing load as much as length. Double stacking certainly won't mean double the boxes per path.

Surely the only limitation on trailing load is the breaking strength of the couplers?
And since we have AAR Knuckles now surely we could make trains as long and heavy as American ones?

Also, I'm not sure how could fit 2 X ISO 9'6" containers in a loading gauge of just over 6metres, especially if OLE is involved.

6150mm is just over 20'2" - which is the loading gauge height defined by the AAR [Plate H] as being suitable for double stack high cube operations.
Well wagons enable the bottom of the container to be extremely close to the top of the rail. On standard gauge it is believed to be impossible to do double stack operations on standard flat wagons as there are serious stability issues if you go much higher.

That does not include 25kV clearances. In fact AAR recommends a structural clearance of over 7.4m for overhead wiring - although the Channel Tunnel shows that is probably very conservative.

Finally, all those new over bridges that have been built over the last 50 years to GB+ would need rebuilding, including every single Motorway bridge. As well as all the other bridges and tunnels of course, and all the OLE (almost all of which would need new masts). Rebuilding tunnels is not easy, as Farnworth, Ipswich and Southampton show us. And the latter two were being given an extra 300mm, not nearly 10 times as much.

The overbridges and tunnels weighed against every platform on the route, a lot of older bridges, probably a lot of the tunnels as well - which is what you need to do anyway the GB+/GC conversion.

Many tunnels, an obvious example being the Stoke Tunnel on the ECML, would likely simply be skylighted to provide the additional clearance.

 

[Bald Rick]

.

Many tunnels, an obvious example being the Stoke Tunnel on the ECML would likely simply be skylighted to provide the additional clearance.

"Simply"

!

 

[AM9]

... So in other words - although the vertical clearance is extreme there are actually relatively few bridges per mile on the network as a whole, and the only real problem is that electrification equipment would have to be rebuilt.
And since the ECML is going to have lots of structures replaced anyway......

But it's just a thought.

Well as you've mentioned the ECML and base your case for it on there being relatively few bridges per mile, just consider the line from Stevenage to Kings Cross. About 30 route miles with more than half the journey in open countryside. Just looking at the four-track mainline (which includes the Hertford loop) there are:

22 road overbridges
at least 16 footbridges
6 rail overbridges including flyovers
20 tunnel bores​

Hardly 'relatively few' and extremely expensive for just one of the mainlines out of London.

 

[class 9]

Surely the only limitation on trailing load is the breaking strength of the couplers?
And since we have AAR Knuckles now surely we could make trains as long and heavy as American ones?

6150mm is just over 20'2" - which is the loading gauge height defined by the AAR [Plate H] as being suitable for double stack high cube operations.
Well wagons enable the bottom of the container to be extremely close to the top of the rail. On standard gauge it is believed to be impossible to do double stack operations on standard flat wagons as there are serious stability issues if you go much higher.

That does not include 25kV clearances. In fact AAR recommends a structural clearance of over 7.4m for overhead wiring - although the Channel Tunnel shows that is probably very conservative.

The overbridges and tunnels weighed against every platform on the route, a lot of older bridges, probably a lot of the tunnels as well - which is what you need to do anyway the GB+/GC conversion.

Many tunnels, an obvious example being the Stoke Tunnel on the ECML, would likely simply be skylighted to provide the additional clearance.

The limiting factor for Intermodals is length not weight, loops & terminals need to be extended.

 

[HSTEd]

Well as you've mentioned the ECML and base your case for it on there being relatively few bridges per mile, just consider the line from Stevenage to Kings Cross. About 30 route miles with more than half the journey in open countryside. Just looking at the four-track mainline (which includes the Hertford loop) there are:

22 road overbridges
at least 16 footbridges
6 rail overbridges including flyovers
20 tunnel bores​

Hardly 'relatively few' and extremely expensive for just one of the mainlines out of London.

I was not actually proposing converting a main line out of London to such a standard, but it could be done - but I am well aware of the cost of attempting something like that.

The foot bridges are not the major issue particularily, they only cost a few million a piece (even with the extra clearance they would likely only cost £10m each or something).

The tunnel bores that have stuff built over them are the major cost.
Especially in the KGX approach.
Lowering the track/tunnels there is likely your only recourse, and probably a pair of locks on the canal north of the station to lift it over the tracks.

However, as people keep suggesting GC/GB+ that would require hugely extensive works as well with even greater operational disruption......

[My initial thought was the axis between Felixstowe and the North West, via places like Ipswich]

EDIT #2:

As far as I can tell the vast majority of the tunnel bores are either in the KGX approach (Gas WOrks and Copenhagen) or between the point where the Hertford Loop diverges heading north, and where it rejoins.

So potentailly you could close that entire section for several months and perform several tunnel reconstructions simultaneously - indeed that also gets you a large part of the route entirely clear for reconstruction of the overhead wire equipment.
You could then close the Hertford Loop for conversion and reopen that section.
It would be horrible for commuters but running maximum length trains on the HErtford Loop at the maximum intensity would help ameliorate it a little.

And Gas Works and Copenhagen are challenging but I believe there are multiple bores, so you might be able to work something out using the reduction of service into the station proper after the THameslink Programme finally gets going.

Or even just run two tracks into maximum gradient descent into a tunnel as soon as possible off the end of the platform at KGX, and just doing a HS1 and running underground the whole way to open country before rejoining the ECML.

"Simply"

!

Well digging it out, considering its relatively shallow and in the middle of the field, is likely to be quicker and simpler than trying to fill it with foam concrete and then turning a TBM loose on it, which is the solution taken at Farnworth.
You would have to prop the tunnel whilst it was being excavated, but not to the extent required there.

Or given that it is in the middle of a field, just building a cutting adjacent to it but far enough away not to cause it to collapse

The limiting factor for Intermodals is length not weight, loops & terminals need to be extended.

It was Bald Rick that suggested that, not me. After all there are far heavier intermodal trains elsewhere in the world.

 

[Shaw S Hunter]

Good grief!!! The most significant lesson learned from the WCRM is that upgrading Victorian infrastructure is horribly expensive for comparatively little gain. If an enlarged loading gauge really is so necessary just build new lines to modern standards: much more cost effective. Unless of course the OP is growing an orchard of fast growing moneytrees in their back garden...

 

[theageofthetra]

Good grief!!! The most significant lesson learned from the WCRM is that upgrading Victorian infrastructure is horribly expensive for comparatively little gain. If an enlarged loading gauge really is so necessary just build new lines to modern standards: much more cost effective. Unless of course the OP is growing an orchard of fast growing moneytrees in their back garden...

Unfortunately the sort of twerps objecting to HS2 lack the knowledge to realise this. Do what China does. Build it & relocation

 

[gimmea50anyday]

This discussion is exactly why we should have built HS2, 3, 4 and 5 30 years ago!

 

[SpacePhoenix]

How many trillions of pounds would it cost to enlarge the loading gauge? It'll never happen due to the cost (which would probably be eye wateringly insane), the time it would take and the disruption that it would cause for probably decades

 

[jopsuk]

So what would you do about passenger stock? If we assume that this involves adopting European standard platforms, existing UK stock will be incompatible. You would need to do something along these lines:
1. Rebuild the platforms
2. Add a temporary surface, similar to the arrangement used in 2012 at Stratford International, to bring the platforms back to current UK standard.
3. Build and fully test a new fleet. Move them, not by rail, to brand new depots and sidings connected to a route.
4. Over a blockade for an entire route, park the entire existing fleet. Remove the platform extensions. Test run the new fleet to make sure the gauging work was OK
5. Reopen and pray the new fleet reliability settles down quickly

Of course,, routes are not isolated. In many places you'd need to retain "classic" platforms in stations for connecting services.

 

[The Ham]

Changing bridge heights is very costly, is not too bad if you are lifting a bridge that is going over a cutting so it has some approach ramps when it didn't before. It gets difficult if you are lifting any other sorry of bridge, as for every 1m of rise the ranks need to be at least 20m longer and the embankments need to be 4m wider.

For a typical road of 6m wide (many are less, but then many are more) which starts 3m above the railway it wood cost something like £500,000 for each side assuming:
- no additional land costs
- no retaining structures
- no junctions within 120m of the bridge
- no utilities (water, gas, electric, etc.) need changing
- no other difficulties

Given that nearly every bridge will have those you are looking at potentially 10's of millions for each bridge and that's before you look at changing the height of the OHLE which would be prohibitively expensive.

As soon as there is any tunnels on the line they would have to be closed for weeks if you are going to rebore them. That on its own would cost a small fortune in rail replacement travel.

Also the rolling stock would be potentially unstable, just anyone that has driven a lorry, van or caravan in high cross winds how much they can get buffeted about. Then add in the cost of a custom train design for a relatively small market and the cost of the trains would be eye watering.

All together it would possibly be not that more expensive to build a new HS line to replace the line. It would almost certainly be quicker and easier.

 

[DarloRich]

cost prohibitive sadly

 

[NotATrainspott]

Any line busy enough to justify this massive expense is too busy for any expense to be worthwhile upgrading it. If you want a double-stack freight route then you're better off doing as the Dutch did and just building one from scratch.

 

[furnessvale]

On standard gauge it is believed to be impossible to do double stack operations on standard flat wagons as there are serious stability issues if you go much higher.

Indian Railways doublestack on standard wagons. Presumably they have NO bridges at all on the routes they do that.

Other than that comment, I see your entire proposal as out of the question due to cost and disruption. I would be more than happy to see ALL freight routes in GB passed for W12 and 775m long trains but I very much doubt I will live to see even that.

 

[coppercapped]

The proposal may be only just a thought, but some consideration has to be given to the economics.

In other words would the increase in receipts over and above those that are already being taken and/or the reduction in operating and maintenance costs of this proposal be sufficient to pay for the work itself, the cost of servicing the debt it incurs and the cost of disruption to the existing services while the work is ongoing?

 

[HSTEd]

So what would you do about passenger stock? If we assume that this involves adopting European standard platforms, existing UK stock will be incompatible. You would need to do something along these lines:
1. Rebuild the platforms
2. Add a temporary surface, similar to the arrangement used in 2012 at Stratford International, to bring the platforms back to current UK standard.
3. Build and fully test a new fleet. Move them, not by rail, to brand new depots and sidings connected to a route.
4. Over a blockade for an entire route, park the entire existing fleet. Remove the platform extensions. Test run the new fleet to make sure the gauging work was OK
5. Reopen and pray the new fleet reliability settles down quickly

Of course,, routes are not isolated. In many places you'd need to retain "classic" platforms in stations for connecting services.

It is most certainly not adopting European Standard platforms, indeed that is the whole point - to take a traditional British loading gauge and merely make it taller

Indian Railways doublestack on standard wagons. Presumably they have NO bridges at all on the routes they do that.

Indian railways also don't use standard gauge

 

[Holly]

I can certainly see how retrofit is economically infeasible in many or most cases.

But new construction is different. As I understand it, HS2 is being built with a loading gauge that fits the needs and provides value and nothing more is economically justified. Today that is.

But since humans are fallible then in 100 years' time the chances that it is exactly correct are slim. It will be too large or too small.
I doubt it will be too large.

 

[squizzler]

Platforms seem to me to be the main perpetrator for gauging woes and also something that is apparently easier to address. If British standard vehicles came specified with retractable platform gap fillers or steps then there could be a gradual switch to set back platforms, either at the TSI height of 760mm or something like 1250mm for level boarding of high floor trains. The benefit is 3+2 seating with seats of the existing width and 25% more capacity as each route is re-platformed.

Don't think it can be done? Caltrain's new Stadler sets are coming with two sets of doors to be able to share with the high platforms on California's HSR. I don't think two doors are needed to transition from UK platform to a TSI platform.

I was reading an Railways Evolution supplement recently in Modern Railways that was talking about gauging stock. I didn't really understand it and was left with the impression that the vehicle gauge is a living thing rather than just whether the cross section drawing of a particular vehicle fits into a particular gauge outline. So there is hope that vehicles can evolve bigger as routes are gradually improved, but the platform is a fixed interface that needs a planned jump to a different standard.

 

[GB]

Timing Load (Trailing Weight) AND length is a factor for freight.

 

[GB]

Couple of points.

Most Intermodal trains are now limited by trailing load as much as length. Double stacking certainly won't mean double the boxes per path.

Also, I'm not sure how could fit 2 X ISO 9'6" containers in a loading gauge of just over 6metres, especially if OLE is involved.

Finally, all those new over bridges that have been built over the last 50 years to GB+ would need rebuilding, including every single Motorway bridge. As well as all the other bridges and tunnels of course, and all the OLE (almost all of which would need new masts). Rebuilding tunnels is not easy, as Farnworth, Ipswich and Southampton show us. And the latter two were being given an extra 300mm, not nearly 10 times as much.

Indeed. And even putting vertical clearance to one side for a moment the OP has not addressed the increase in axle weight over the track, culverts and other bits of infrastructure as far as I can see.

 

[RichmondCommu]

Surely the only limitation on trailing load is the breaking strength of the couplers?
And since we have AAR Knuckles now surely we could make trains as long and heavy as American ones?

.

Surely one of the key constraints has to be the length of refuge loops?

 

[HSTEd]

Surely one of the key constraints has to be the length of refuge loops?

Well yes, but you can run an end ring around such things using various methods that increase the weight of the train per unit length - like double stacking containers.

Changing bridge heights is very costly, is not too bad if you are lifting a bridge that is going over a cutting so it has some approach ramps when it didn't before. It gets difficult if you are lifting any other sorry of bridge, as for every 1m of rise the ranks need to be at least 20m longer and the embankments need to be 4m wider.

For a typical road of 6m wide (many are less, but then many are more) which starts 3m above the railway it wood cost something like £500,000 for each side assuming:
- no additional land costs
- no retaining structures
- no junctions within 120m of the bridge
- no utilities (water, gas, electric, etc.) need changing
- no other difficulties

Given that nearly every bridge will have those you are looking at potentially 10's of millions for each bridge

So peanuts in the grand scheme of things?
We are talking about multi billion pound schemes.


And as a counterpoint to the counts of overbridges and tunnels on the section between Stevenage and KGX, between Stevenage and Peterborough there are:
- 34 Road Overbridges
- 1 Rail Overbridge
- 15 Foot Overbridges

Over 54 miles or so, so much further than between KGX and Stevenage.

and that's before you look at changing the height of the OHLE which would be prohibitively expensive.

The clearance between the old wire and the new wire would be so massive that you could probably assemble the new overheads above the old ones during closures, complete with registration and such, then gas-axe down the entire original installation.

The cost would probably not be that much more than electrification of the original line was - whilst it would be more difficult working around the existing equipment, you can reuse most of the non track infrastructure, such as substations and bulk supply points and such.

Also the rolling stock would be potentially unstable, just anyone that has driven a lorry, van or caravan in high cross winds how much they can get buffeted about. Then add in the cost of a custom train design for a relatively small market and the cost of the trains would be eye watering.

As opposed to our endless succession of custom trains for the relatively small boom-and-bust UK market?
And there are no recorded major stability issue with double stack container operations in the US, despite it being rather widespread.
It might be tall but remember it is also quite heavy as well.

All together it would possibly be not that more expensive to build a new HS line to replace the line. It would almost certainly be quicker and easier.

~£30bn and 15 years of construction, based on the example of HS2?
That sounds rather excessive considering you yourself say tunnel conversions would only take weeks.

 

[RichmondCommu]

Well yes, but you can run an end ring around such things using various methods that increase the weight of the train per unit length - like double stacking containers.

However you then have the problem of tunnels and bridges and their height restrictions.

 

[Chester1]

Am I correct in thinking that new bridges and tunnels built to W12 gauge? The age of our railway network and current growth indicates that infrastructure built now may very likely be in use in a hundred years time. It seems short sighted that we haven't been building to UIC GC gauge (excluding platforms) when it is affordable. It would at least mean no future worries about the gauge of freight trains. The pressumption could be GC gauge with NR needing to apply for exemptions for any new infrastructure planned that will not be built to fit GC gauge trains.

 

[Tio Terry]

Well yes, but you can run an end ring around such things using various methods that increase the weight of the train per unit length - like double stacking containers.

What would that do to axle loading and the problems that brings with things like track design, under bridges and culverts, stress on S&C, the need for more speed restrictions and the effect that makes to train timings and other services on the same route? You cant just effectively double the axle load without having to pay for it. Then there is the question of rolling stock design to be able to deal with the higher weights and side on wind loadings.

 

[AndrewE]

Indian Railways doublestack on standard wagons. Presumably they have NO bridges at all on the routes they do that.
.

Could the fact that their "Standard" gauge is 5'6" have something to do with it?

 

[Grumpy]

One thing that I have not often heard suggested, is that we could develop a specialist loading gauge that would be identical to the British loading gauge as far high as the cantrail, and would then simply continue upwards for a much larger distance, likely terminating with a flat roof more compatible with modern rolling stock design practices.

.

An interesting suggestion.
We have a number of routes where passenger trains are at capacity and it is not clear how continuing growth can be accommodated with existing infrastructure. Building brand new railways may be a cheap solution in rural areas where you are just paying the cost of acquiring farmland but it simply wont be possible or economic in and around the major cities. People seem happy to quote WCML when arguing that you cant upgrade an existing railway, but seem to forget that the ECML was upgraded efficiently with minimal disruption. And also that Railtrack was a managerial shambles.

So consider lines south of the Thames with no OHLE. Put freight to one side as a red herring and reconsider whether you really need 6.1m height.

 

[Bald Rick]

A few more thoughts. The question isn't:

"Would it make sense for the country to go for a very high gauge", but is:

"Would it be better to go for a very high gauge rather than a 'standard' European gauge"

All hypothetical of course, as neither are going to happen, because neither make sense.

However, to answer the exam question, the answer is, possibly, yes, albeit perhaps not for the reasons many have stated.


Firstly, the point of going for a very high gauge with existing W6 gauge below solebar level is to avoid working on platforms and other low obstructions to European gauge.

It therefore comes down to would it be cheaper to rebuild every tunnel / OLE system and almost every bridge (for very high gauge) than to rebuild a smaller proportion of them but have to rebuild every platform and low obstruction for European gauge. The answer to that would be different for every line.

However one thing is necessary to point out. The main European loading gauges (Universal, GA, GB, GB+, GC) are wider across the whole width of the train, and not just below the solebar.

This means that when going to any of these European gauges there are far more difficult things than platrorms and low obstructions. Firstly, almost all tracks in this country would be too close together. They would need separating. Almost every junction would need modifying, as the fouling points would change, as well as the track separation. Complex junctions - such as at busy stations and freight yards (where double deck trains would be most likely to operate) would need wholesale remodelling. It cost hundreds of thousands of pounds just to modify a short stretch of track at Ripple Lane to enable continental gauge wagons to access a certain siding from HS1 for example.

Moving the tracks apart also means many retaining walls, fences, cutting walls, signals, and most OLE masts would be in the way, at least on one side. This is in addition to bridges and tunnels, which would need rebuilding. So, Alterations required to most civil engineering, almost all the track, stations, signalling and OLE. It's effectively rebuilding the railway.

So for this reason, I think that yes, it would probably be cheaper to go very high rather than standard European gauge.


To answer the implied question of "in that case does it make sense" the question is a comprehensive "Non"

It would, still, require complete replacement of all the contact systems elements of overhead electrification, almost all bridges and tunnels, and many signals. I can think of no route where this would be even remotely cost effective. Tunnels for example - it is simply not the case that you can lower the floor of a tunnel by 2 metres. For most tunnels, once you go below about 500mm you hit the invert which is a structurally integral part. That means you need to rebore it. Opening out tunnels isn't exactly straightforward either. The big cost in tunnelling today is not the TBM, or the folks working it, but getting rid of the material that comes out. Opening out a tunnel to the required width for stable cutting slope angles produces a phenomenal amount of material - much much more than a tunnel of the equivalent length. It also requires a lot of land, which gets expensive if there is anything other than fields on top.

The number of high density passenger flows that have sufficient demand and capacity issues that could require double deck services that coincide with high levels of intermodal flows that could do with double stacked freights is pretty limited. Basically the GEML from Ipswich to Stratford, the NLL east of Willesden, the WCML south of Rugby, and Southampton to Basingstoke. Naturally you would have to clear to very high gauge for the whole service route for it to work, and that is where problems arise - the approaches to Liverpool St for example, where you can't go down, there's tunnels in the way, so you would have to demolish the £2bn of real estate on top. Or Hampstead Tunnel. Or Watford Tunnel. Etc.

Finally, the point about overturning. It is true that double stacked container trains run in the US, in well wagons with a low centre of gravity. Nevertheless I have heard (and seen!) plenty of examples of them coming off in high winds - but on US Railroads it rarely matters - they are almost all single track, and largely in the middle of nowhere. That doesn't apply here, and remember we have had boxes blown off standard container flats on the WCML relatively recently.

Passenger vehicles of the same height would, however, not have such a low centre of gravity. And I suspect that at slow speed on highly canted track, with some unfortunately loading (A sumo team all on one side of the top deck) and the wind the wrong way would produce a rather unfortunate result.