Oil Train

[Wizlish]

Euclid

I wonder what engineering reasoning drives the design of this type of tank car, which is apparently used in Russia. It looks modern enough, but it is surprising that they use a separate frame with the tank lashed into place with straps. I wonder if those are loose wooden cushions that are laid into the tank saddles and cradle the tank.

I think this is a cryogenic car, perhaps a LNG car.  The straps hold the tank in place with minimum contact area, and the padding strips in the 'saddle' brackets would be some kind of foam.  I think the 'smooth' outside is the result of extensive lagging with a jacket over it, and that the visible end shield may have (insulated) transfer piping behind it.

What did Globaltrans say about the car's construction or use?

Compare the VTG design here, which also uses careful methods of insulation and 'tank suspension' to cut down on losses from cryogenic contents::

http://www.railwayage.com/index.php/mechanical/freight-cars/vtg-unveils-europes-first-lng-tank-car.html?channel=59

[daveklepper]

Coupler problems.    Have not mu car operations pretty much solved this problem?  I am not referring to subway systems, where the Westinghouse coupler with its built-in electrical and air connections is pretty standard, but mus such as Silverliners and the old Metroliners that had to be compatible with railroad standard Janney-derived couplers and still handle multiple air and electrical connenections.  Were there problems with the Metroliner couplers?  I don't remember any, but that might be because of the news of other greater problems.

Metro North and LIRR mus have a non-compatible coupler, however.

[CSSHEGEWISCH]

IC mu cars, both the original fleet and the Highliners, used Tomlinson couplers.

[Euclid]

Here is an interesting tank car [from previous page]:

http://www.globaltrans.com/uploads/pics/rail-rank-car_01.jpg

 

Wizlish,

I think you could be right about it being a cryogenic car.  That would explain the seeming isolation between the tank and the frame as the purpose being for thermal isolation.  I can’t find any specific information from Globaltrans. 

[Euclid]

https://www.aar.org/Fact%20Sheets/Safety/CBR%20Fact%20Sheet_2_8_15.pdf

Quote from the above link by the AAR:

“Braking Systems – Railroads are equipping all trains with 20 or more carloads of crude oil with either distributed power or two-way telemetry end-of-train devices. These technologies allow train crews to apply emergency brakes from both ends of the train in order to stop the train faster.”

 

QUESTION: Specifically how does add safety to oil trains?

[oltmannd]

Euclid

https://www.aar.org/Fact%20Sheets/Safety/CBR%20Fact%20Sheet_2_8_15.pdf

Quote from the above link by the AAR:

“Braking Systems – Railroads are equipping all trains with 20 or more carloads of crude oil with either distributed power or two-way telemetry end-of-train devices. These technologies allow train crews to apply emergency brakes from both ends of the train in order to stop the train faster.”

 

QUESTION: Specifically how does add safety to oil trains?

 

Two ways  One is it gives another way to initiate an emergency brake application.  This would be useful if an anglecock handle got turned somewhere mid-train.  

The other way is it would start the emergency brake signal pressure wave from both ends of the train, taking a few seconds off the time to get the whole train in emergency, reducing stopping distance.

[Euclid]

Thanks Don,

I understand your first point.  But the second point raises a couple questions:

Am I correct in assuming that this quicker stopping would only apply to an emergency application initiated by a crew member, and not to one initiated by a derailment causing air hoses to part?

How much quicker would an emergency application stop the train if it were initiated from both ends of the train as opposed to just the head end?

[BaltACD]

With the two way EOT - if either the engine or the EOT get pneumatic notification of a emergency application - that fact is radioed to the other end and the train gets the emergency brakes applied from both ends.  If the UDE originated in the middle of the train the pneumatic notification would probably reach both the engine and the EOT at nominally the same time.

[Euclid]

If an emergency application were initiated by a mid-train derailment, how does either the head end or the EOT receive notification of the mid-train emergency application initiation? 

If both ends receive this notification via the propagation of the emergency application as it moves away from the mid-train derailment, wouldn't the emergency applications initiated from both ends of the train be too late to do any good?  

[BaltACD]

Euclid

If an emergency application were initiated by a mid-train derailment, how does either the head end or the EOT receive notification of the mid-train emergency application initiation? 

If both ends receive this notification via the propagation of the emergency application as it moves away from the mid-train derailment, wouldn't the emergency applications initiated from both ends of the train be too late to do any good?

Even if it is propagated instantly - it is still too late to do much 'good'.

[Wizlish]

Euclid

If an emergency application were initiated by a mid-train derailment, how does either the head end or the EOT receive notification of the mid-train emergency application initiation?

You have a couple of things in here conflated that have to be separated to make full sense.

The way the derailment is 'signalled' to the brake system is critical.  Presumably what you mean is that Spanish system that triggers an emergency application when a truck frame dips or a bolster rotates a certain amount. 

Two things happen in emergency -- the pressure in the trainline starts to fall as air blow out, and a shockwave pulse is sent down the trainine.  The latter happens both ways from the 'point of incidence', and the emergency valves on the cars trigger as this signal reaches them.  So the signal will propagate pneumatically to the closer of the two 'ends'.  That end will then 'radio' the emergency to the other end and put that valve in emergency, in turn sending another shockwave back down the line to meet the one propagating from the detector.

As you can imagine, if the derailment occurs precisely midtrain, there would be no change in braking effect at all; the emergency valves will all have been triggered by the time the sonic signal reaches the valves at the ends.  If the derailment is closer and closer to one end, the benefit of triggering the 'other' end promptly begins to increase, becoming essentially the simultaneous front-and-rear triggering for derailments in the first or last cars.

What needs to be looked at here is the effect on train dynamics when the brakes are applied 'at sonic speed' spreading out from multiple points like this.  For all intents and purposes a 'derailed' car acts just like a dynamiter, and there will be a spreading patch of deceleration in the middle of the train centered around the physical point of derailment, lagging the sonic application speed by whatever the delays built into the brake-valve response, cylinder application time, etc. produce.

If the on-car detector sends radio to the valves at head and tail instead, the reaction will be exactly as for a standard emergency application initiated from the head end and mirrored by the EOT.  Naturally, if the train has distributed "EOT-like valves" every block or x number of cars, you'll get pressure waves propagating midtrain, forward and back from each valve location, and  very quickly this matches the application state that would be achieved in full ECP with emergency application on a per-car basis commanded nearly simultaneously (say, about .93c, which I recall being about right for DC in an insulated copper conductor)

If the "signal" for the derailment is a parted air hose, the above bets are off.  Note that at least one coupling probably has to part for the hose to do so, so we're probably at the 'digging-in' phase of derailment mechanics; the shock then propagates in both directions from the site of the separation until one end is reached, then the radio initiates a pulse from the other end promptly.  Again, I can see complex interactions in train dynamics occurring from these patterns of brake being set, compared with the ECP (or multiple-EOTlike-valve) approach.

[Euclid]

Wizlish,

I was not talking about derailment sensors, but I can see how that would come to mind in my question about the signal propagation.  I asked the question just to make sure of my assumption that signal would propagate to the head end and EOT by dynamiting reaction through the brake pipe.  I could not see any other way it could happen, but if that is how the signal travels, I was thinking that it would to be too late to contribute to speeding up the emergency application.  But then Balt explained that whichever end receives the brake pipe signal first radios the opposite end to open the brake pipe there.  So then I realized that it could speed up the application to the extent that the derailment happened nearer to one end or the other.  And as you say, if the derailment happened in the exact middle of the train, there would be no benefit in the action of the distributed power or two-way telemetry end-of-train devices. 

I can see your point if two-way telemetry end-of-train devices were combined with derailment sensors, and if there were multiple versions of the EOT device dispersed throughout the train, that would begin to match the performance of ECP

However, as I recall, ECP speed of application is not just due to its simultaneous call for application, but also in its lack of flow restrictors that are required for conventional air brakes. 

So altogether, considering that ECP stops only about 3% quicker than conventional air brakes in the “Emergency” application; the improvement offered by opening two holes in the brake pipe rather than one would seem to be quite small at best, and non-existent in some cases.     

[oltmannd]

Euclid

Thanks Don,

I understand your first point.  But the second point raises a couple questions:

Am I correct in assuming that this quicker stopping would only apply to an emergency application initiated by a crew member, and not to one initiated by a derailment causing air hoses to part?

How much quicker would an emergency application stop the train if it were initiated from both ends of the train as opposed to just the head end?

 

You were asking about the AAR statement..."These technologies allow train crews to apply emergency brakes from both ends of the train in order to stop the train faster.”

 "How much quicker" is going to depend on train size and crew reaction, but likely diference would be on the order of few seconds.  I'll leave the kinematics to you.

[Wizlish]

Euclid

Am I correct in assuming that this quicker stopping would only apply to an emergency application initiated by a crew member, and not to one initiated by a derailment causing air hoses to part?

Don's answer was correct, but terse enough that you may not comprehend everything he said.

The technology of the 'EOT mirroring' provides that as soon as a brake valve is put in emergency at the head end, the rear end follows suit.  That says nothing about how the valve came to be put in emergency in the first place.

In a derailment, there will be more than one associated 'reaction time' from the crew before the brake valve moves to emergency and the rear end actuates.  There will be a fixed time for the crew to recognize there is an 'anomalous event' at all (and as noted in recent cases of derailment, this time may be very long or even indefinite).  Then there is decision about how to react -- the level of braking that should be applied.  Then there will be the physical time to reach to the brake nandle and move it to where the emergency application will start to propagate.

If all these reactions have taken place before a derailment has progressed to the point where an air hose parts, the emergency application will progress more quickly even net of all delays, and will do so as 'both ends toward the middle' with associated kinematic effects on the train 'as a whole' and on the derailed section, however that section may be 'evolving' with time.

If these reactions have not completed when an air hose parts, you will have a spreading 'island' of emergency appllication, proceeding at the effective speed of the signal in both directions, with the associated lags at each car for the valve to react, and then for the air to reach the cylinders, take up the slack, etc., followed by application of both ends (1) if the crew recognizes something is amiss and puts the head-end valve in emergency before the signal wave reaches either end, or (2) as previously described when the 'island' reaches whichever end is closer, and the other end responds.

In the case of the Spanish-style derailment detector, the midtrain emergency application occurs with no more latency than the physical derailment kinetics (e.g., truck rotation).  I think it is highly likely that the resulting 'island' of emergency signaling will have propagated to the 'nearer' end of the consist by the time the crew reaction times have gone to the point the head-end valve is opened enough to have signalled the rear end valve to actuate.  So net of human reaction times the strictly-pneumatic derailment detector may give the shortest achievable setup to emergency, and by extension the quickest possible emergency application, that can be achieved with single-pipe triple-valve equipment.

As Don said, whether or not the kinematics in the train are favorable or unfavorable to the derailed cars in that scenario is something you will have to work out for yourself, in a process that will involve at least two steps: (1) figure out the effective braking effort car-by-car over time (as the signal propagates and the lagged responses occur) and then (2) translate this using the same timescale into decelerations of the cars and forces between them. 

Now, stepping back and taking a breath for a moment, my own preference for how a pneumatic 'derailment detector' would work is just a little different: it would produce a modulated response in the service range to a derailment event, and it would be set up to give stronger braking in proportion to the severity of the condition that is detected.  This has the propagation speed in the train of a service application, and it is probably not going to propagate a proportional response 'from one end to the other' because it involves a pressure difference, not detection of the 'edge' of a sonic signal.  What would be required would be for the two 'ends' to mirror even comparatively small but monotone-decreasing pressure differences ... this becoming a potential issue for consistent leaks that cause the rear-end pressure to be arbitrarily lower than front-end pressure while the train is in service.  Engineering decisions would follow (like, for example, does detection of a 'monotone-decreasing pressure' at one end trigger a proportional rate of decrease at the other end, or do you promptly 'match pressures' so the other end initially sets up starting at a somewhat higher rate?)

(I find I have made a couple of ASSumption, without realizing it, which are (a) that the method used to apply 'the other end' simultaneously involves recognition of the pressure signal in the air line itself, and not the physical position of the brake handle or part of the valve, and (b) that the method involves pressure state as well as the 'signal' that triggers emergency applications.  A separate detector that senses both would seem be the logical way to implement things for DPU, but stranger things have happened with aerospace engineers 'optimizing' railroad design...  you should confirm how the actual 'emergency' signal is sent, detected, and implemented, and there are people on list in a poaition to confirm exactly how it is done for all locomotive designs.)

[Euclid]

Wizlish

 

Euclid

Am I correct in assuming that this quicker stopping would only apply to an emergency application initiated by a crew member, and not to one initiated by a derailment causing air hoses to part?

Don's answer was correct, but terse enough that you may not comprehend everything he said.

No, Don’s response was not too terse for me.  As I mentioned, when he gave that answer, I was picturing the signal traveling to both ends, but arriving too late to quicken the application due to the fact that the travel of the signal is caused by the propagation of application.  So it follows that the signal reaches the ends of the train after the application is complete.

That is why I asked Don this: 

“Am I correct in assuming that this quicker stopping would only apply to an emergency application initiated by a crew member, and not to one initiated by a derailment causing air hoses to part?”

Then in the next post, Balt explained that the first end of the train to receive the signal will immediately transmit it to the opposite end, thereby eliminated any time remaining for the signal to travel to that end by the pneumatic propagation; that is, to the extent that the derailment is not in the middle of the train.

In my question, I made the distinction between an application initiated by a crewmember as opposed to one initiated by a derailment.  This is because I assumed that an application by a crewmember would initiate at both ends simultaneously. 

[Euclid]

oltmannd

 

Euclid

Thanks Don,

I understand your first point.  But the second point raises a couple questions:

Am I correct in assuming that this quicker stopping would only apply to an emergency application initiated by a crew member, and not to one initiated by a derailment causing air hoses to part?

How much quicker would an emergency application stop the train if it were initiated from both ends of the train as opposed to just the head end?

 

 

 

 

You were asking about the AAR statement..."These technologies allow train crews to apply emergency brakes from both ends of the train in order to stop the train faster.”

 "How much quicker" is going to depend on train size and crew reaction, but likely diference would be on the order of few seconds.  I'll leave the kinematics to you.

 

A few seconds sounds about right if the application is made by a crewmember.  However, if the application is initiated by an air hose parting during a derailment, the effect of distributed power or two-way telemetry end-of-train devices would not offer any advantage if the derailment occurred in the middle of the train, and would only approach the “few seconds” advantage to the extent that the derailment occurred nearer to either end of the train. 

So the only certainty of receiving the full advantage of distributed power or two-way telemetry end-of-train devices would in in cases where a crewmember applied the brakes.  This is by far most likely when the purpose is to stop in time to avoid a collision. 

I can see two distinct ways in which quicker stopping offers advantages in improving oil train safety:

1)   Providing more time to avoid a collision.

2)   Reducing the time length of the derailment process and the amount of damage that it causes.

 

Only item #1 could consistently provide the full advantage of distributed power or two-way telemetry end-of-train devices.  Item #2 will offer that advantage in amounts that range from zero to full, depending on where the derailment occurs within the train. 

[Wizlish]

Euclid

So the only certainty of receiving the full advantage of distributed power or two-way telemetry end-of-train devices would in in cases where a crewmember applied the brakes. This is by far most likely when the purpose is to stop in time to avoid a collision.

Which was (and is) the principal reason for this design, and for the 'advantages' of ECP's more rapid application in service braking (and slight advantage in emergency braking).

The picture has now changed with 'stopping safely after a derailment' becoming the principal situation of interest.  I expect it will be a while longer before the stakeholders work through just how different that situation is.

I can see two distinct ways in which quicker stopping offers advantages in improving oil train safety: 1) Providing more time to avoid a collision. 2) Reducing the time length of the derailment process and the amount of damage that it causes. Only item #1 could consistently provide the full advantage of distributed power or two-way telemetry end-of-train devices. Item #2 will offer that advantage in amounts that range from zero to full, depending on where the derailment occurs within the train.

You're still conflating two things that you ought to keep separate.  One is the speed with which the brakes are commanded to apply, and the other is the speed with which they actually produce progressive retardation of the train.  (A third thing, which is tied in with all these discussions of train dynamics, is the degree to which, and the locations in which, the braking creates or affects longitudinal forces in the train, which may affect the dynamics of a derailed car or cars). 

I have not yet seen a definitive technical accounting -- buslist, and perhaps others, can easily provide one -- of what is in that "3% difference" in emergency stopping distance between ECP and single-pipe brake systems.  I have been assuming that there is more to this than the quicker transfer of air to effective brake=cylinder force that is possible for ECP valves that do not 'dynamite', but it is also possible (the magnitude is about right) that we are looking at the difference between 'near-immediate' setup of emergency actuation on all cars vs. progressive sonic actuation from the control valve(s), with the 97% reflecting the actual time that the brakes proceed to apply and set and mechanically stop the train.  One thing I'd like to see as a table or graph is the emergency response from (1) a conventional head-end emergency application; (2) a head-and-tail emergency application with propagation from both ends, (3) a couple of examples where an emergency release in the train propagates through the consist, and (4) full electric actuation of emergency valves on each car in the train near-simultaneously.  This might be plotted as bars left-to-right with the length either being time or distance (with the time taken to set up the brake to a common degree of actuation, I suggest when the brakeshoes first touch the treads, being one color, and the actual stop after that point being another).

That at a glance will indicate the magnitude of the effective differences between the approaches, provide a proportioning of the time and distance that's attributable to the signaling and control technologies,  and show where the practical improvements might be made.  I believe tables of data (from WABCO) have already been posted that show some of the data, but in a different format.

Something else that would be of potential advantage would be to graph the time and/or distance for various train speeds, as the proportion of time involved in setup will be constant independent of speed, and the distance traveled during the setup will of course be directly proportional to the speed, but the physical braking distance will be a predominantly factor of the square of the speed. 

(As an aside, it might be a useful finger exercise or BS discriminant to see what speed corresponds to the nominal "3%" improvement, depending on what assumptions were made ... including what the 3% is an improvement in, stop time or distance, or application time...)

[Euclid]

Wizlish,

You cite these two things that you insist I am conflating (quoted from you):

“One is the speed with which the brakes are commanded to apply, and the other is the speed with which they actually produce progressive retardation of the train.”

************************************

Yes, I absolutely understand that those two things are separate, and they influence the stopping distance independently.  But where did I ever say otherwise or conflate them, as you say?

You say that I am conflating those two things when I said this:

I can see two distinct ways in which quicker stopping offers advantages in improving oil train safety:

1)   Providing more time to avoid a collision.

2)   Reducing the time length of the derailment process and the amount of damage that it causes.

 

The two things that you say I am conflating are two independent components of stopping speed.  Yet my comment in blue has nothing to do with comparisons of stopping speed.  Instead, it is about two independent objectives of improved stopping speed.  So, in my comment in blue, there is no reason to draw a distinction between the two things that you say should not be conflated. 

Regarding my comment in blue, this is my point:

I would conclude that item #2 is far more important because it will apply to every derailment, no matter what the cause.  Derailments caused by collision will be far fewer than those due to other causes.  Derailments caused be a collision that could have been avoided by a couple seconds less stopping time will be far fewer yet.   

[Wizlish]

Euclid

1) Providing more time to avoid a collision.

2) Reducing the time length of the derailment process and the amount of damage that it causes.

Regarding my comment in blue, this is my point: I would conclude that item #2 is far more important because it will apply to every derailment, no matter what the cause. Derailments caused by collision will be far fewer than those due to other causes. Derailments caused be a collision that could have been avoided by a couple seconds less stopping time will be far fewer yet.

All that was established a very, very long time ago.  Collisions have little, if anything, to do with oil-train safety.  So it is not much of a 'conclusion' -- the only reason I started discussing collisions at all is that you seemed to be mentioning them with respect to brake performance in some way.  (I do think it is important to mention at this point that I agree with the premise that in very few HHFT accidents will a brake application initiated by the crew with the intent to stop the train as fast as possible, as if reacting to an imminent collision, turn out to 'make a material difference' in safety of outcome.)

Now please tell just how 'far more important' you now think your point #2 is in absolute terms.  Because this was a point Dave Husman was trying to establish a while back, and you got into something of an argument with him about it. 

 

[Euclid]

Wizlish,

I agree that preventing collisions is relatively unimportant.  Part of my point in bringing up collisions (objective #1) is exactly to de-emphasize their importance in relation to objective #2.  When I say that objective #2 is far more important, it is because mitigating the damage of derailments by shortening their duration is far more likely to prevent damage, fire, death, and injury than stopping just in time to avoid a collision.  Of course, avoiding a collision will prevent a derailment which is better than mitigating it.  But it will be so infrequent that the mitigation will be of more benefit overall.

The other part of my point is that the AAR preferred method of improving braking offers the largest benefit to point #1(the least important objective) and less to point #2 (the most important objective).

I am not entirely sure what Dave Husman’s position is on the objective of reducing the time length of the derailment process.  I gather that he thinks it is completely unattainable to any extent.  Others here have made the same point.  In my opinion, the AAR has implied it.  I think this is somewhat of an industry paradigm.  That is, when the first wheelset leaves the rails, nothing can be done to change the course of action or mitigate the damage.  I disagree with that premise.  I think that the ECP technology opens the door to advanced concepts in derailment mitigation.  In that sense, ECP is a technology in its infancy.   

[Wizlish]

Euclid

The other part of my point is that the AAR preferred method of improving braking offers the largest benefit to point #1(the least important objective) and less to point #2 (the most important objective).

This is one of the most important points to stress, but I am not sure you want to tar the AAR alone with the brush -- I would add how simple (in principle) it is to augment the performance of the 'two-valve' AAR system with something -- and the Spanish have demonstrated it can be a simple and relatively reliable something -- that signals the air when a derailment is detected.

I'm surprised that Dave Klepper for one has not observed that, once you have ubiquitous DPU radio (or equivalently, ubiquitous transception between front end and EOT) it becomes very simple to install pressure sensors in the trainline every few cars, or in fact in every car as it comes in to be shopped.  The only thing this has to do is transmit a coded signal when it detects a pressure event.

There is now only x cars' worth of propagation before both ends set up, even with no intermediate valves every so often in the train.  I would suspect that this, as an HHFT 'mandate', is orders of magnitude cheaper than compatible ECP, and offers almost the whole of the prospective benefits in reduced stopping time  and distance.

I continue to think that the future for derailments lies in more careful reduction of momentum rather than 'crash stops' (what a double entendre, that!), and in particular with some form of differential braking.  This will require much more careful design, and then much more careful framing of the discussion about HHFT accident causes and what to do about them, specifically including the difference between brake systems optimized around avoiding collisions ... which, remember, is a principal premise behind PTC and its mandate ... rather than handling a derailed or unstable train with maximum care.

In brief, in my not-so-humble opinion, the whole idea of slamming on the brakes as a reaction when wheels aren't on the rail to do any braking is a bit misdirected to begin with.  What needs to be overcome is the perception that a bad alternative rigidly enforced is superior to a better alternative that might from time to time produce a worse result.  The lawyers, of course, will argue the thing whichever way they see pay to do.

[Euclid]

Wizlish

I continue to think that the future for derailments lies in more careful reduction of momentum rather than 'crash stops' (what a double entendre, that!), and in particular with some form of differential braking.  This will require much more careful design, and then much more careful framing of the discussion about HHFT accident causes and what to do about them, specifically including the difference between brake systems optimized around avoiding collisions ... which, remember, is a principal premise behind PTC and its mandate ... rather than handling a derailed or unstable train with maximum care.

In brief, in my not-so-humble opinion, the whole idea of slamming on the brakes as a reaction when wheels aren't on the rail to do any braking is a bit misdirected to begin with. 

Yes, absolutely.

Starting with a train equipped with ECP brakes, equip every car in the train with a derailment sensor for each truck.  Use a mechanical sensor such as the one developed in Spain.  Then have the sensor/detectors communicate signals to the ECP system upon detection of a derailment.  The response from the ECP system would be tailored to account for a variety of variables including train speed, tonnage, power level, any dynamic braking underway, any ECP service braking underway, slack position, and location of the derailment in the train consist. 

With these factors, the ECP system would determine and apply a specialized brake response that would gently stretch the train into and through the derailment zone as it develops.  It would divide the train into two independent braking ranges consisting of the cars ahead of the point of derailment and the cars behind that point.  The braking force in these two ranges would be different.  That is the “differential” in differential braking.

The objective of this specialized ECP response would be to dissipate kinetic energy of the train as much as possible without having the derailment turn into a pileup.  At the same time the ECP system begins to implement this specialized application, it will adjust power, adjust any ECP service braking that happens to be underway, and take the train out of dynamic braking if it is applied.   

Most specifically, this system would not respond to a derailment with a normal “Emergency” application of the brakes as it done with the Spanish system.  I believe the “Emergency” response could immediately cause a pileup in a derailed and dragging train that otherwise might just continue to drag further.  The longer it drags without piling up, the more rolling energy is dissipated.  The more rolling energy that is dissipated during the derailed-dragging phase, the less there will be to inflict damage if the cars do pile up.   

[dehusman]

Focusing on the braking systems and train dynamics to prevent pileups only works if the braking systems and train dynamics determine whether there is a pileup.

If you read the AAR statistics and actually look at the actual causes of the oil train derailments with pileups it becomes pretty evident that the braking systems and the train dynamics had very little to do with there being a pileup.  It depends on what caused the derailment in the first place.

The accidents where the track was obstructed or the track/roadbed was the cause ended up piling up.  If you have a big hole in the roadbed or something fouling the track, then no amount of differential braking or slack control will help the situation.  The only thing that will help the situation is stopping the train. 

[tree68]

Indeed - gentle braking will only help with gentle derailments - ie, those where the train is able to continue, such as dropping a truck onto the ballast.

As soon as anything causes an abrupt stop (ie, that truck turns and digs in, hits a switch, finds a defect in the track structure) all the braking in the world isn't going to make any difference.

Go back and look at the tornado video.  As soon as the first car left the tracks, the train was in emergency - not gentle braking - and still the rear portion of the train ran into the locomotives at a good clip.  

[Euclid]

tree68

Indeed - gentle braking will only help with gentle derailments - ie, those where the train is able to continue, such as dropping a truck onto the ballast.

As soon as anything causes an abrupt stop (ie, that truck turns and digs in, hits a switch, finds a defect in the track structure) all the braking in the world isn't going to make any difference.

Go back and look at the tornado video.  As soon as the first car left the tracks, the train was in emergency - not gentle braking - and still the rear portion of the train ran into the locomotives at a good clip.  

 

The differential braking system that I have described will create tension by pulling slightly into and through the derailing cars to keep them stretched and sliding.  This would be done as an automatic programed function of the ECP system.  It would not accomplish this by increasing the braking behind the derailment.  The cars behind the derailment would receive maximum braking force as they do now.  The tension would be created by reducing the brake force on the cars ahead of the derailment.  So it has to hold off some of the braking on the engines and the cars ahead of the derailment.   

What happened in the tornado video would not happen with differential braking because the engines would not be allowed to exercise their full stopping power as they did in the video.   In order for differential braking to work, cars and engines ahead of the derailment must have their braking reduced below that of the cars behind the derailment.   

So the effect in the tornado train video, if it had differential braking, would have turned out basically opposite to what actually happened.  With differential braking, the engines would have kept rolling, so the cars behind would have had time to stop without the tank car catching up with the engines and hitting them.

The differential braking system would have made sure that the trailing cars did not catch up with the leading cars.  That would be its basic function by nature in providing stretch. 

In some cases, the pulling into the derailed cars could begin consuming more energy than the freer rolling leading cars can provide if there are not enough of them, such as when a derailment happens say 5-10 cars behind the engines.  Even if the system left no braking on the 5-10 cars ahead of the derailment, it might not render enough kinetic energy to keep the derailing cars stretched if there were many of them.  In those circumstances, the differential braking system will call for power from the locomotive to add to the kinetic energy of the free rolling head end. 

So this system of differential braking will tend to be gentle enough to encourage a single derailed-dragging car to not start a pileup.  But it is not limited to do that with just a single car.  It could work on a dozen derailed cars all coupled together, and sliding at high speed like a person on water skis.  Cars trailing may eventually pileup before it all gets stopped, but a dozen cars skidded clear of the wreck ahead of it is a dozen cars that don’t get into a highly destructive pileup.    

Actually, the tornado video shows this water ski effect quite vividly in the moment after the first car rolls over and opens a line of sight.  You can see at least a couple cars on their sides and moving along in a straight line, coupled together, and still moving quite fast.  

[tree68]

OK, fine.  Now, break an axle, have it dig in, and tell us how differential braking is going to save the day.  Remember, the car with the broken axle just went from 40 to zero MPH in about 40 feet.

[BaltACD]

^

^total diavowal of train dynamics and not worth the bytes that have been harmed in formulating it.

[Euclid]

tree68

Indeed - gentle braking will only help with gentle derailments - ie, those where the train is able to continue, such as dropping a truck onto the ballast. 

I want to clarify something.  It sounds like you are referring to “gentle braking” to mean that the entire brake application is made at a very low braking force. 

Above, I said this:

“With these factors, the ECP system would determine and apply a specialized brake response that would gently stretch the train into and through the derailment zone as it develops.”

When I said, “gently” stretch, I was referring only to the stretching being gentle so as to do some good, but not pull the train in two.  The overall braking effort will not be reduced in order to make to it gentle.  That overall braking effort will only be slightly less than with conventional air brakes.  That will come from using slightly less brake force on the train ahead of the derailment.  Then that slight difference in braking between the head and hind cars will cause the “gentle” stretch.   

[Euclid]

tree68

OK, fine.  Now, break an axle, have it dig in, and tell us how differential braking is going to save the day.  Remember, the car with the broken axle just went from 40 to zero MPH in about 40 feet.

 

It won’t go from 40 to zero MPH in 40 feet.  It will simply leave its broken axle behind if the car remains coupled and under tension.  Look at the Lynchburg wreck.  Eight cars “dug in” and left their trucks behind in pieces, but the cars kept going without them. 

[tree68]

Euclid

It won’t go from 40 to zero MPH in 40 feet.  It will simply leave its broken axle behind if the car remains coupled and under tension.

I should have been more specific - the truck digs in, after the wheel falls away....