Bad FRA Regulations
Since many people are linking to my previous post identifying the FRA as the primary obstacle to an American railroad revival, I’m hoisting a comment I wrote on the Infrastructurist detailing some of the FRA regulations that are the most destructive.
The original references for this are from Zierke’s website and the East Bay Bicycle Coalition, but those are a few years out of date, and recently the FRA has made noises about reforming the first two rules, which are the most destructive to intercity rail. Unfortunately, those reforms are not good enough, chiefly because they are designed to preserve the FRA’s bureaucracy, piling more obstacles on any attempt to modernize US trains.
1. 945 tons buff strength for locomotives and end cars and 360 for coaches (link); the maximum that’s even partly defensible is Europe’s 200, and Japan’s 100 is perfectly safe. This is by far the most important: as a result of this rule, the Acela power cars weigh 90 metric tons, vs. 68 for the TGV power cars they’re derived from. Zierke notes that the lighter the train, the higher the FRA weight penalty is.
2. 4″ maximum cant deficiency for non-tilting trains, except 5″ on track connected to 110+ mph rail (derisively called the magic HSR waiver by railfans). The Acela is limited to 7″ despite tilting. Non-tilting TGVs do 180 mm in France (about the same as the Acela) and tilting trains do 250-300 mm in Japan and a bunch of European countries, no special testing required except on actual track. In addition, superelevation is limited by regulation to 7″ minus a safety margin; high-speed lines around the world have 180 mm actual superelevation, and the Tokaido Shinkansen, which has tighter curves, has 200 mm.
Those two regulations are already being somewhat modified. Amtrak seems to believe that the nationwide mandate for positive train control (PTC), passed in 2008 in response to the Chatsworth crash, will allow it to run lighter trains; the FRA has granted Caltrain a waiver from the FRA buff strength rule provisioned upon PTC installation. As for cant deficiency, the FRA has already decided on a revision allowing tilting trains up to 225 mm cant deficiency, and non-tilting trains up to 150 mm by testing.
Unfortunately, those two reforms only look good at first glance. The Caltrain waiver application from the buff strength rule was devised in consultation with the biggest rolling stock manufacturers – Bombardier, Kawasaki, Alstom, and Siemens – which indicates which rules they could comply with and which they could not. This may well lock out smaller vendors, such as Stadler and CAF. Stadler’s FLIRT is the fastest-accelerating, highest-powered regional train on the market; it is also very light, and may well not comply even with regulations Caltrain did not ask out of.
In addition, since such waivers depend on PTC, if the freight railroads succeed in their attempt to delay or water down PTC implementation, which they consider too expensive, then future rolling stock purchases will remain heavy. Indeed, Amtrak’s purchase of new electric locomotives, due to enter service in 2013, is FRA-compliant and more expensive than purchases of similar locomotives in Europe; this despite the fact that they are intended to run on the Northeast Corridor, which has a PTC system.
As for the cant deficiency waiver, it was obtained by testing existing outdated technology in the US, such as Amtrak locomotives and the EMUs used on commuter rail in the Northeast. No attempt was made to use high-cant deficiency European technology, a point also made by Drunk Engineer. Such trains would have to be tested to the FRA’s satisfaction, and not be allowed to run at the same speeds as they do in Europe. In fact the FRA’s proposed rule revision includes a language about higher track standards for cant deficiency higher than 5″, never mind that TGVs run on less than perfect legacy track at 7″ cant deficiency.
In addition, for high-cant deficiency operation, it’s important to regulate both cant deficiency and the rate at which it changes. The muscles can adjust to lateral acceleration, given enough time; thus the jerk, or the rate of change of acceleration, must also be prescribed. With a proper superelevation ramp and change in cant deficiency based on the abilities of existing trains, high speeds and high cant deficiencies can combine well, as found in a Swedish study about the feasibility of very high-speed trains on legacy track.
Additional FRA regulations, which hamper regional rail more than intercity rail, seem to be here to stay. These include the following:
3. Two employees per train; regional trains should have one. But, bear in mind, many regional operators have multiple conductors, and the limit to lower staffing is antiquated trains or managerial incompetence rather than the FRA. For example, the MBTA believes it needs one conductor per two cars.
4. Brake tests at every turnaround. Intercity trains can enter a stub-end station and back away in 3-4 minutes, and do every day in Germany; regional trains turn around in 3-4 minutes in Japan. However, Amtrak makes Keystone trains dwell 10 minutes at Philadelphia.
5. Four-quadrant gates required for quiet zones; these make quiet zones expensive, and as a result trains have to blare loud horns at grade crossings, alienating neighbors and creating NIMBYism.
6. No regulations encouraging high-performance lightweight cars and good signaling. The FRA should mandate a modern system, preferably ETCS, which permits a throughput of up to 37 trains per hour at standard speeds. This is 12 tph more than currently can run between New Jersey and New York, and would be about $13 billion cheaper than Amtrak’s Gateway tunnel proposal, which would add 21 tph.
The multitude of bad regulations is why I think FRA reform has to be intensive, without any half-measures. The new rail regulations in the US should as much as possible be based on UIC (predominantly European) and Japanese regulations, with the present status quo ignored.
The only role of American regulators should be to devise a coherent system to allow European and Japanese trains to interact with each other. In some places, such as PTC and jerk, it requires greater regulation, based on best industry practices in the rest of the developed world. But in most other areas, the rule should as far as possible be that everything that’s legal in Europe or Japan is legal in the US.
I’ll repeat my exhortation in my post on Mica’s privatization plan: please contact the relevant Congressional representatives and let them know that any real reform must include extensive FRA reform. Organization and electronics should come before concrete, and such deregulation of rolling stock could jive well with the conservative mood in Congress that Mica is channeling. And if it does not, then never mind – the Democrats could seize FRA reform, too, as a good-government issue. It’s more important than whether future railroads are run publicly or privately.
Pedestrian Observations 
Theres something about American government that it wants to invent things itself, always, instead of looking to see what works elsewhere. See: healthcare reform.
Quick question – does the two employees per train requirment include or exclude the driver? Either way, it’s a silly regulation… two employees is not sufficient for a long sleeper train, but excessive for a 2-car regional train. The requirment should be that the operator sets out what the minimum safe staffing level should be and why, and the FRA sign off on it (if that).
What I would like to see is a European-style carriag and an FRA-compliant carriage tested to see what kind of impacts they can suffer on different parts of the body without deformation. Ultimately, that’s what matters. (The abilty to withstand head-on collisions only ever mattrs to the front carriage anyway. All the others will get hit at some sort of angle).
It includes the driver. On freight operators, as well as some of the more forward-thinking commuter rail operators, this means an engineer and a conductor.
Caltrain did some simulations of the type you propose. Although FRA-compliant cars can survive bigger impacts without deformation, they are less likely to survive impact without danger to the passengers, which is ultimately what should matter. FRA compliance turns out to only protect the lead car in a very narrow set of situations, namely a head-on collision at relative speeds between (if I remember correctly) 15 and 25 mph. Crumple zones are much safer than buff strength; lighter trains are safer at grade crossing accidents; lighter trains also derail much less at equal cant deficiency.
VIA rail also has this stupid requirement; it will happily stack people in two cars of a 3 car trains because since a train fire some years ago, it will insist that every occupied car shall have one employee (even just a peanut-seller).
In North America, it seems that the travelling public is treated as drooling morons, as exemplified with the zeal tickets are checked prior to boarding a train in big cities (as much as 4 times in Montréal); in Europe, where millions of people pass through stations every day, such a scheme is certainly not feasible. But in Europe, they have extensive signalling and passenger information.
But there is fortunately some change; ever since the revival of commuter trains in Montréal, the AMT no longer checks tickets at the gates, and even allows people on terminal platforms before the train has arrived!
Wow. Is it really that bad over there? In the US they do let you on the platform before a commuter train arrives, though not before an Amtrak train arrives. And they check tickets only on the train, and a few forward-thinking operators even do POP, though they still retain a conductor for regulatory reasons (and in Caltrain’s case assistant conductors for historical reasons…).
Does what you describe come from federal regulations, provincial regulations, or just internal agency incompetence?
I don’t think it’s regulations… Toronto commuter rail operator GO Transit has one driver and one other staff member on its trains (12-car double deckers carrying ~2,000 people). GO also does POP with uninhibited platform access…. although at Union station, Toronto, they tend to not announce the platform until after the previous train has cleared, to reduce crowding on the narrow platforms. The contrast between VIA and GO’s operations is very noticable at Union.
Probably because of liability… See my post below regarding the ex-Nightstar rolling-stock.
Montréal-area commuter trains (as well as GO) are now POP systems. As it happens, in Québec, the relevant laws were changed some years ago, and the STM (Montréal city transit) has also changed to POP, although there are turnstiles in the Métro, and you have to board the front door and confirm you have a valid transfer/pass by putting your pass on the contactless reader (so called because you have to rest it for 1 second against the reader itself before it registers).
That goes a long way to explain why VAIO has such a high number of employees (3000+) for its ridership (4.2m). Related issue: at stations, they have one employee by every door to help passengers down onto the platform (which is probaly why coaches only have doors at one end). The better option would be to ask passengers “will you need help borading/alighting” when they buy their ticket, and put those passengers in the same coach… that way only one helper is needed.
NB: I did a quick compare of passengers per employee with VIA and all the UK rail franchises (because I can find the numbers easily). Commuter-orientated franchises come out higher (better), inter-city franchises come out lower. The two lowest were 9,000 and 6,000 passengers per employee, while VIA manages only a measly 1,400.
Employees have been calling the “new” ex-Nightstar trainsets (dubbed “renaissance”) “déplaisance” (french for “displeasure”), as they are notoriously manpower-hungry in Montréal and Québec city.
Those are the two only high-platform stations in Canada, and due to the small british loading-gauge of the ex-Nightstar stock, there is a ≈30 cm gap between the door and the platform, which has to be bridged by a small portable platform, and someone has to be attending to make sure no one falls upon the track. Hence the high manpower “requirement”.
I’m getting tired of that Flirt hype. Compare it to DBAG 423: – slightly older train tailored to high-platform Münchner S-Bahn needs and with direct competitor, Coradia Continental:
Power arrangement: Flirt’s power modules occupy space for passengers, DB 423 and C.C. have their power modules under floor and roof-mounted respectively
Powered axles: Flirt has just two powered trucks out of typically 5, limiting it’s acceleration rates in low-adhesion conditions, the other trains have just one unpowered truck out of 5
Specific weight: 1.72/1.55/2.00 t/m – high floor is the biggest saver here, C.C. suffers because of roof-mounted traction.
Specific hourly power: 21.7(26.0) / 22.4 / 20.57 kW/t
For me, 423 is the clear winner, followed by Coradia Continental (because of it’s better adhesion and better floor-space usage that may correct it’s per-passenger values).
I’m getting tired of that Flirt hype. Compare it to DBAG 423
Isn’t it a bit difficult to compare the FLIRTs to the S-Bahn stock? The former basically destroys anything for regional railway purposes, while the latter is a high-density suburban subway?
No, the DBAG 423 is comparable in power – actually somewhat higher-powered, because it doesn’t have the FLIRT’s low-floor weight penalty. It’s rated at 140 km/h because it’s used on S-Bahn lines that don’t need to run at 160 km/h, but if they did, the train could probably be easily rated at higher speed.
If so, admittedly leaves one questioning why the FLIRT stock is capable of higher acceleration when compared to the ET 423*? One would think that the higher acceleration would be needed on that rolling stock…
*Admittedly, I’ve been on both, and they’re still better than anything that operates around New York.
It’s higher initial acceleration, which is a question of adhesion – for example, rubber-tired equipment is quite good at initial acceleration. But beyond a certain point, it’s limited strictly by power – namely, at speed v m/s, a train with k kW/t is limited to power providing k/v m/s^2 of acceleration.
> Isn’t it a bit difficult to compare the FLIRTs to the S-Bahn stock? The former basically destroys anything for regional railway purposes, while the latter is a high-density suburban subway?
No, it is not, because FLIRT is much more typical S-Bahn stock than DBAG 423 (because most of cities didn’t bother to raise platform height of their S-Bahn stops from 0,38/0,55/0,76 m to at least 0,96 m, killing compatibility with long-distance rolling stock).
> If so, admittedly leaves one questioning why the FLIRT stock is capable of higher acceleration when compared to the ET 423*?
That’s the point I was trying to make. Top acceleration is weight on drivers times adhesion factor divided by train weight. Weight on drivers and train weight are “hard” numbers that don’t change much, but the adhesion factor is highly variable, depending mostly on track and wheel conditions (wet leaves can drop it to 1/4 of value on dry rail), followed by powertrain type and then powertrain control. Given FLIRT’s comparatively low weight on drivers, it’s reliable acceleration – the one needed for schedule creation – is actually lower than reliable acceleration of both class 423 or Coradia Continental.
In Prague, they learnt the importance of adhesive weight the hard way, when replacement of ČD classes 451/2 with class 471 lead to slower running times, because class 471 (with 35-40 % of weight on drivers) couldn’t keep up with class 451 acceleration (around 60 % of weight of driver) despite class 471 superior specific power and powertrain type. (technical data: http://bit.ly/kz1q9j , http://bit.ly/j5MiAQ ).
Does anybody actually have acceleration profiles for any of these fast EMUs? It’d be nice to make timetables for lines drawn on maps (or rather, stations drawn on existing commuter lines).
There are YouTubes of the FLIRT’s acceleration (it loses 24 seconds accelerating to 160 km/h, and 13 accelerating to 100 km/h), and Clem posted a link to a KISS acceleration curve in a comment on CAHSRBlog in early May.
For Japanese trains, you can get decent estimates by comparing local with express timetables – just be careful to make sure the local you look at doesn’t get held for a few more minutes to let an express train overtake it. On the Chuo Line, between Nakano and Tachikawa, where the top speed is 95 km/h and the tracks are dead straight, local trains lose about a minute per stop – a bit more at rush hour, a bit less in the off-peak.
Just to be clear: when you say ‘looses 13 seconds accelerating to 100 km/h’ you mean that by the time the train reaches 100km/h, it has spent t number of seconds and driven a distance d, whereas with 100km/h will run a distance d in t – 13s?
Because then one could just calculate with 1minute delay for a stop in a ‘slow’ stretch, and 1.5min in a fast one.
Yes, this is exactly what I mean. The deceleration penalty from 100 km/h is about 10-11 seconds. With 30-second dwells, it works out to about 55 seconds per stop at 100 km/h, and 75 at 160.
“The FRA should mandate a modern system, preferably ETCS, which permits a throughput of up to
37 trains per hour at standard speeds. ”
Alon,
That ETCS system is talking through put on mainline tracks. That’s not taking into consideration that Penn Station isn’t mainline tracks. The current HDIS system installed in the North River tunnels is capable of handling 30 trains per hour. The limiting factor isn’t the signaling system. So neither HDIS, ETCS, nor any other system is going to help.
The limiting factor is how many trains can be moved through A interlocking on the west side of Penn Station and how many trains can be handled on Penn’s 21 tracks while safely loading/unloading the all the passengers. Amtrak is already taking some steps to improve A interlocking, however Gateway would see a radical change to A interlocking removing a considerable bottleneck.
The LIRR & Amtrak can put many more trains through to Penn from the east even without HDIS, in part because of the extra 2 tunnels, but in large part because there are two interlockings on the east end of Penn, Jo & C interlocking. This greatly reduces the number of conflicting moves during busy times.
Penn’s 21 tracks are not a real capacity limit. The operators think they are because they are used to very long turnaround times, but in reality I’ve seen LIRR trains close doors within less than 2 minutes of opening them, at the height of rush hour. If trains run through, or otherwise terminate and turn quickly (which means FRA regulations about brake tests need to go), then Penn can see its number of tracks halved and still not have any capacity problems.
If they insist on grade-separating Harold, which allows conflict-free movements from Hell Gate to the southern tunnels, then it’s exceptionally easy to just use tracks 10-13 for through-service; if they see reason and spend the money elsewhere, they can at vastly lower cost construct such switches that tracks 14-19 will be the through-tracks without any single-track segments (it’s currently possible using existing infrastructure as long as tracks 16 and 17 aren’t used by trains going in opposite directions).
I would like to see a 1920’s era PRR/LIRR employee timetable to see how many trains per hour were operated during that time… I somewhat suspect that it was far more than 30 per hour…
Where can we read this in English without so many undefined acronyms? Don’t you want normal people to be able to understand what you are talking about? If you only want people already intimate with these concepts to be able to understand you, you are mostly preaching to your own choir.