Oil Train

[oltmannd]

Euclid

Incidentally, I did try to call New York Air Brake to ask if empty/loaded sensors are being applied to new tank cars as a matter of course.  But, although you can pick up the phone and call them, you can’t expect anyone to answer.  I have not seen any photographs of tank cars where the load sensors are evident. 

Probably not equipped since they have relatively low load/tare ratio and can keep the loaded braking ratio high enough without it.

[Euclid]

In studying this drone video, I have developed the following analysis of the oil train derailment in Lynchburg, VA last year.

Refer to this video: https://www.youtube.com/watch?v=j8kONLIWlBs

 

LYNCHBURG, VA OIL TRAIN DERAILMENT MAY 2, 2014

 

Prior to the emergency brake application, and prior to any separation of the train; the train derailed at least eight cars in succession with each derailing at the initial point of derailment for the first derailed car. 

By the time these eight derailed cars stopped, the first car to derail was approximately 1,200 feet past the point of derailment.         

The eight cars remained coupled together, mostly upright, and in line with the track bed, despite the fact that they destroyed the track and lost all of their trucks in the dragging process.

Derailment Sequence of events:

1)    First car (GTCX 742045) derails, establishes point of derailment (P.O.D.), and begins dragging and damaging track.  First car is designated as car #1 in this analysis.

 

2)    Car #2, following car #1, derails at P.O.D. due to the track damage cause by car #1.

 

3)    Cars #3-8 derail in succession at P.O.D. with each dragging car increasing the track damage.

 

4)    Cars #1-8 are all dragging in line with the track bed while the entire train is still coupled together, and no emergency brake application has been initiated.

 

5)    Car #9 derails at P.O.D. and the train separates between cars #8 and #9.

 

6)    Upon separation, the air hoses between cars #8 and #9 separate, causing the entire train to initiate an emergency brake application.

 

7)    As the brakes begin to apply, the portions of the train ahead of and behind the separation continue moving with more events unfolding within both portions of the separated train.

 

8)    The portion ahead of the separation of cars #8 and #9 continues to drag derailed cars #1-8.

 

9)    After car #8 moves approximately 200 feet past the P.O.D., the forward portion of the train separates between dragging cars #5 and #6.

 

10)The portion ahead of the separation of cars #5 and #6 continues to drag cars #1-5.

 

11)After car #5 moves approximately 60 feet past the separation of cars #5 and #6, the forward portion of the train separates between dragging cars #1 and #2.

 

12)The portion ahead of the separation of cars #1 and #2 continues to drag car #1.

 

13)After car #1 moves approximately 1,200 feet past the P.O.D., the forward portion of the train stops with car #1 still coupled to the car ahead of it.  Note: The trailing truck of the car ahead of car #1 also derailed at some point in this dragging process.

 

14)As the train separates at the P.O.D, between cars #8 and #9, and the emergency brake application begins, the portion behind the P.O.D. continues moving forward, led by the derailed car #9.

 

15)Because derailed car #9 has lost the pull of the cars ahead of it upon separation, and because it is being shoved ahead by the cars behind it, car #9 has no guidance whatsoever.  So it makes an angular excursion from the track line.

 

16)As a reaction to car #9 making an angular excursion, the car immediately behind it makes an angular excursion from the track line in the opposite direction, thus beginning a jackknifing progression that involves the next eight cars including car #9.

 

17)Rear portion of train stops.

[dehusman]

You assume that the force that was keeping the train from jacknifing was the pull of the engines on the front.  The train derailed, the train parted and the lack of pulling allowed the train to jacknife.

I believe the events happened in a different order.  The train derailed, the train jacknifed, the train parted and the head end had nothing to do with the jacknifing.

[Euclid]

Dave,

I agree that the train might have jackknifed before parting, as you say.

I am not necessarily assuming that what prevented jackknifing for a time was the pull of the engines, although that is one possibility.  But it also might have been partly or entirely pulled by the kinetic energy of the train portion ahead of the derailment pulling against the mounting resistance of the derailing and dragging cars.  In any case, the derailing cars would be like brakes applying to cars in the middle of a train.

If that mounting resistance of the derailing cars began to significantly decelerate the train, it may very well be that the cars behind the derailing cars shoved into them with their own kinetic energy and shoved hard enough to buckle the train and begin the jackknifing.  In that case, the jackknifing occurred, and was immediately followed by the parting between the first car to jackknife and the car ahead of it, as you suggest. 

However, I tend to think that is unlikely for this reason:  Every car ahead of that jackknifing would have been slack-stretched, including the eight derailed cars and all the cars still on the rails ahead of them.  Therefore, if the trailing cars were to shove against the derailing cars, where would the resistance strong enough to cause jackknifing come from? 

I suppose the trailing cars could have momentarily bunched the slack in the derailed cars (and cars ahead of them) and created enough resistance in them to cause the jackknifing which in turn caused the parting.  And then after the parting, the slack ran back out in the derailed cars as they continued dragging.

That would explain the scenario that you describe where the jackknifing occurs before the parting.  In my scenario, the parting occurs before the jackknifing.  In your scenario, the jackknifing occurs due to buff force in the un-parted train, and then causes the parting.  In my scenario, the parting occurs due to the track reaching a sufficient degree of destruction, and then causes the jackknifing. 

In either scenario, the parting and start of jackknifing are nearly simultaneous. And no matter which scenario is true, it does not affect the main point I am making about the derailment.  That point is that eight cars derailed and did not get involved in a pileup.  As such, those eight cars did not breach or burn.  By the time those eight derailed cars stopped, the lead derailed car had moved 1,200 feet past the point of derailment, and did so before the jackknifing began and before an emergency brake application began.  That is a lot of lost time in reacting to the derailment with a brake application.

Had this train been equipped with derailment sensors, a brake application would have begun before cars #2-8 derailed and began dragging, with their lead car traveling 1,200 feet past the point of derailment.

The beneficial effect of dragging the first eight derailed cars to safety ahead of the ensuing pileup, in this case, was happenstance.  With ECP differential braking, the same effect would occur by design.  And in that case, there would have been immediate braking behind the point of derailment which would have begun reducing the ability of the trailing cars to shove into the derailing cars and buckle the train.  With derailment sensors and ECP differential braking, this derailment might have ended with no pileup, no breaching, and no fire.

You say that the “train parted and the head end had nothing to do with the jackknifing.”  I agree with that.  But the head end had everything to do with dragging eight cars clear of the jackknifing.  That is the point that I am making.  

The counterpoint that others have made here is that once the first wheel leaves the rail, it begins a chaotic process that is completely out of control.  I disagree with that.   

[tree68]

So, how fast was the Lynchburg train going when the derailment occurred? (It's probably been stated, but to review...)

Euclid

Therefore, if the trailing cars were to shove against the derailing cars, where would the resistance strong enough to cause jackknifing come from? 

That force would come from the first car that jacknifed going from X MPH to Zero MPH in about two seconds flat (or less).  The reason for the jacknife in the first place would be because the car dug into the ballast.  The faster it's going, the harder it's going to dig in.

The head end stopped when it did because the coefficient of friction between the cars and the ballast was much higher than what the brakes would have provided.  

1200 feet is less than a quarter mile - peanuts when you're talking about a loaded, 100 car (more or less) train.  

Once again - go back and look at the tornado derailment.  About six cars were blown off the track, amounting to some 300 feet.  I would opine that it's unlikely that even if there were 1200 feet between the locomotive and the cars still on the rail at the point that the derailed cars were clear of the rail, that the cars still on the rail would have stopped in time.  A derailment detector would have done nothing in that incident.

The remaining cars would clearly have been in emergency (or at least beginning the application), yet look at the speed with which the trailing cars collided with the locomotive.

Had that been an oil train, there would likely have been spills and a fire.

[ruderunner]

So I've been following this thread and see all the complaints about Euclid not letting up on his ideas. All the railroaders seem to be stuck in their ways and pooh pooh his ideas. I'd like to point out that the same reactions were spouted about steam engines an flanged wheels.

I see that Euclid s ideas do add up and can make a tremendous improvement in train handling, not just derailments. But it's not something that will happen soon. Given time for ecp to mature, and for it to be added to the fleet pushes his complete system out at least 10 years.

That said how about a simpler approach?  But first some questions, can a FRED dynamite the brakes?  Do they have valves that can?  How difficult would it be to enable such a feature and have it be cab controlled?

I'm seeing this as a spinoff of the differential braking idea that Euclid has. If the engineer senses a derailment (meaning wheels offtrack but not to the point of jacknife) he can start the braking from the rear and keep power applied up front to stretch the train.

Iirc cabooses had such a valve that the conductor could activate.

The way I see it modified FFREDs and programming to support such a feature is small potatoes in cost and certainly worth trying.

And why aren't load/empty valves standard? Many pickup trucks had them prior to ABS brakes and they worked. Its a proven technology that wouldn't be hard to adapt to the fleet when cars are in the shop.

I see these 2 systems being easy enough and inexpensive enough to add to trains especially unit trains with their dedicatd routes and consists.

Re ECP and derailmentsenfutureleaving costs out of the discussion since we already know they are expensive to implement, it seems the 2 biggest problems to solve would be power supply and communication. Its been stated that adding the necessary wiring would be frought with connection troubles between cars. I agree. So what about on board power and wireless communication?  The transmitters need not be powerful a range of 100 feet would be adequate if the transmitters could also repeat the signal and send it down the length of the train. This should be weak enough that the FCC shouldn't have too much trouble accepting it.

Power supply could be in the form of an axle mounted generator and battery. But again nothing huge like needed on passenger equipment. Automotive sized components should be adequate and inexpensive.

Like I said this isn't in the near future but I believe it is coming...

[dehusman]

ruderunner

So I've been following this thread and see all the complaints about Euclid not letting up on his ideas. All the railroaders seem to be stuck in their ways and pooh pooh his ideas. I'd like to point out that the same reactions were spouted about steam engines an flanged wheels.

There are a lot of things that Euclid brings up that would be really cool to have.  The problem is that there is a degree of difficulty in establishing them that Euclid doesn't seem to understand.  So many of his ideas take off on wild tangents with things he doesn't understand or physical actions he ignores or refuses to accept.

Remember Euclid said that "big things were coming" in the oil train standards and it would incluld a heavily armored car and ECP and new couplers, etc, etc, etc.  The professional railroaders said that the new car standards would be an evolution of the standards for flammable gas cars.  Have you read the recommendations?  They are an evolution of the flammable gas cars.

I see that Euclid s ideas do add up and can make a tremendous improvement in train handling, not just derailments. But it's not something that will happen soon. Given time for ecp to mature, and for it to be added to the fleet pushes his complete system out at least 10 years.

It will take 10 years at a cost of billions or trillions of dollars.  You should really read some of the literature.  Here's an interesting statistic from a Federal report.  If you converted 98% of all the rail cars in the US to ECP, the chances of having a 100 car train with all ECP is less than 40%.  How many short lines and small shippers will you put out of business or drive away from rail because its not economical to spend a million dollars on upgrading their car fleet?  Abstract questions to railfans, real problems to real railroaders.

But first some questions, can a FRED dynamite the brakes?  Do they have valves that can?  How difficult would it be to enable such a feature and have it be cab controlled?

You are a couple decades late to the party, EOT's that can initiate an emergency application have been around for a long time and have been required on trains operating on substantial grades for years.

 

One thing all the studies seem to ignore and there has been very little discussion about is DPU.  It can also intitiate an emergency application.  It also mimics all the the brake activity (sets, releases) that the head end engine does.  DPU cuts the reaction time by the factor of at least how many DPU sets there are (one rear DPU cuts the application time in half, mid and rear cuts it about a quarter to third).

 If the engineer senses a derailment (meaning wheels offtrack but not to the point of jacknife) he can start the braking from the rear and keep power applied up front to stretch the train.

And how to you think he can sense the derailment? If he's dragging a car he might feel the drag increase a bit but that's about it.  Chances are the first time he knows its derailed is when it goes in emergency.

The way I see it modified FFREDs and programming to support such a feature is small potatoes in cost and certainly worth trying.

So did the railroads, that's why they did that 10-20 years ago.

And why aren't load/empty valves standard? Many pickup trucks had them prior to ABS brakes and they worked. Its a proven technology that wouldn't be hard to adapt to the fleet when cars are in the shop.

And that's the trap.  "Proven" technology is not necessarily reliable technology.  Most ECP applications have been in unit trains.  You make all the electrical connections and they stay together for months.  What happens when you put them in general freight service and couple and uncouple them a dozen times each loaded trip?  What happens when you uncouple them and leave the connections exposed to the elements for a month, two months, a year, then try and use them.  The trials the railroads have had haven't been that favorable.  If the railroads have a good system that will work 99.9999% of the time and better system that works 98% of the time they will go with the good system.  The operating costs associated with the extra one or two percent less reliability would be millions or billions of dollars in  delay costs.

I see these 2 systems being easy enough and inexpensive enough to add to trains especially unit trains with their dedicatd routes and consists.

That's because you don't see the additonal operating costs (bad orders, cars used in other than unit service, cars used in other service, interchange, etc).

So what about on board power and wireless communication?  The transmitters need not be powerful a range of 100 feet would be adequate if the transmitters could also repeat the signal and send it down the length of the train. This should be weak enough that the FCC shouldn't have too much trouble accepting it.

If that happens then the failure of one car would immobilize the entire train.  You now have to rely on  150+ power supplies, 150+ transmitters and 150+ recievers to work.  Impossible?  No.  Jacks up the degree of difficulty?  Yes.  The railroad environment is harsher than possibly any environment except the military.  The equipment has to function at 120 degr, -50 degr, bone dry desert,  monsoon rains.  They have to be able to sit for 6 months then fire up immediately and work flawlessly non-stop for years.  No updating the software once a week, no running dianostics, no boot errors, no software crashes.

Power supply could be in the form of an axle mounted generator and battery. But again nothing huge like needed on passenger equipment. Automotive sized components should be adequate and inexpensive.

Now you will need to replace batteries every couple years, so now you need stores of batteries at every repair location.  What happens when a car doesn't get used for 6 months (not uncommon) and the battery dies/loses it charge/goes bad during that time?  How do you move the car (or cars since there are thousands of cars that fall into this category)?  If a wheel goes bad, if its the one with the axle mounted generator, what happens if the repair point doesn't have that type of wheel set?  what happens if the wheel goes bad out in the middle of nowhere?

(Technically I'm just jerking your chain, they would probably use an air operated turbine-generator set like they have been using on EOT's for the last decade or so, as soon as the train gets air on it, it has power, it just needs a battery for very short periods of time.  But that illustrates why we get frustrated, 75% of the "innovations" are really old news, been there, tried that. )

 

[Euclid]

Dave,

You have mentioned several times the flammable gas cars as being a model for the more crash resistance oil cars.  But you have never made it clear what the connection is.  I don’t feel like doing deep research to connect the dots on how the flammable gas car standards are going to be applied to oil cars. 

If the flammable gas car standards are going to fix the problem with oil cars, why aren’t they being applied now?  Why have we not heard what they will entail?  Why did the industry decided on building 1232 cars as the solution to the problem?  Why did the industry miscalculate on the performance of the 1232 cars? 

What is the range of load weights for a railcar load of flammable gas?  My guess is that it is a lot lighter than a load of crude oil.  So is it possible that the flammable gas cars can be strengthened adequately without compromising their load capacity whereas oil cars cannot?

I did propose a safe oil train concept, as you say, but it never included “a heavily armored car” as you say.  My concept is in lieu of strengthening tank cars, as opposed to including that approach.  I also never said "big things were coming" in the oil train standards in reference to my proposal.  I only proposed it as a solution, but never predicted that it would be adopted.  Of course it won’t be adopted.  It is too much too soon for such a massive bureaucracy, which is generally too big and too standardized to accept sweeping change.  It took the industry over 60 years to solve the “waste grab” problem in journal boxes by using a woven pad instead of loose cotton waste.  And by then, waste lubrication was obsolete due to roller bearings.  If anything is changed in oil train practice, it will be by government mandate.

I agree with your points about the practical problems with ECP brakes.  Because of those problems, I do not believe that the universal conversion will ever happen.  So when it comes to ECP brakes on oil trains, if follows that it must be limited only to oil trains.  Once that is accepted, it opens the door to further refinements such as empty/loaded sensors, derailment sensors, differential braking, etc. 

Here is my prediction:  The industry will not voluntarily adopt ECP brakes for oil trains.  That alone is too much.  However, ECP brakes for oil trains will be mandated by the government as part of the impending new tank car rules.  So ECP will be added to oil trains as the “overlay” option that offers the flexibility of either brake system, leaving the cars compatible with non-ECP equipped cars.  Oil trains will continue to operate loose cars, so the ECP connectors will be retained.  They may have to be made more reliable, but even though the oil trains will be loose cars, the ECP connectors may be disconnected and reconnected relatively infrequently.  So, I assume they can be made reliable enough.      

[ruderunner]

Dhuesman I realize that none of this is cheap or easy. But we have to start somewhere. I'm just getting frustrated by the old school thinking of  that's not how we do it or was tried 50 years ago and didn't work so it will never work etc etc etc.

Instead of putting the ideas down, I'd like to see some ow could we make it work feedback.

As someone in the auto industry, I can recall when computer controls started. At first they were troublesome but became more refined and reliable. The early systems in the late 50s only controlled fuel and disappeared quickly, next came ignition control, then a combination. Now there isn't much in a car that isn't compute controlled.

Someone somewhere developed manufacturing processesthat allowed for those improvements 

In the railroad world individual roads or other companies are developing their own systems apparently without much thought to interoperability amongst the national network. I'd like to see the roads try to develop something as a team. Preferably before the government steps in with another mandate like PTC.

[zugmann]

ruderunner

I'm just getting frustrated by the old school thinking of that's not how we do it or was tried 50 years ago and didn't work so it will never work etc etc etc.

Fair enough.  But there's also some of us that get frustrated at new school thinking that we must take this "whiz-bang electronic, computerized untested and unproven theoretical device" to modify the most basics of train operations (like brakes).  Also adds much expense and will make the cars that much harder to repair/maintain.

I wonder if the lack of ECP development isn't in due part to newer AC power with extended-range dynamic brakes that can grind just about any train to a halt on all but severe grades?

 

[ruderunner]

Moving on, thanks for the clarification on how capable FRED is. I should have guessed it was tried al tree already. And pointing out the DPU angle.

As for how an engineer could sense a derailment well isn't that what derailment sensors are for? Or perhaps its time to bring back rear end crews?  I know the snowball is getting bigger...

It seems that my comments on load/empty valves are crossed up with ecp. Load/empty can be fully mechanical and reliable. Its also car centric not caring if its loose car or unit train. And to me proven means functional and reliable.

The transmitter problem may not be as bad as you think. It can be setup such that a few failed transmitters won't disable the train. That would be a programming issue. Even looking at a dead transmitter between t good ones, I picked 100 feet as a number that could jump the distance of a car without a transmitter/broken transmitter. If that needs to be raised to 200 fett that's still not a large,powerful,or expensive unit. And the temperature angle doesn't fly for me. Many modern electronics are on daily use under such temperature ranges. Automobile computers deal with this everyday all over the world. Sure vibration from rough track or hard coupling is more severe but a cushioned or sprung mounting would be easy to design and implement.

Air generator, that's a better idea than axle mounted and the kind of feedback I think we need to share. Working together to solve the problems instead of bickering over them.

 

[ruderunner]

And fair enough as well Zugman. But you sound just like mechanics from the early 80s when computer controls  started to appear. But the industry adapted to the technology and now its second nature as well as far less expensive. For reference personal computers cost hundreds of dollars when they can out and weren't very powerful. Today the price is the same but when adjusted far inflation the cost is half or less for computers that are a 100 times more powerful.

I think we need to look past the hhuge initial costs of some of these technologies and look forward to how the costs drop per unit as the number of units increase. Outfitting one car today with all the bells and whistles can indeed run to 1,000,000 but a large run of them may make the added cost only a couple thousand per car. Think mass production versus custom built.

Load empty sensors can be purely mechanical and self contained. They have been around for a long time already I'd just like to see more implementation of them.

Are dynamic brakes a substitute for ECP? I dont think so. Dynamic brakes dont have the future capabilities that ECP can offer. Things like built in load/empty capability, Euclid's differential braking idea, or even antilock braking. Practical now? No but sometime in the future it may be.

[tree68]

ruderunner

As for how an engineer could sense a derailment well isn't that what derailment sensors are for? Or perhaps its time to bring back rear end crews?  I know the snowball is getting bigger...

Reference the story in trains some time back (and a more recent incident I know of but won't get into), where a derailed car ran for several miles and then re-railed itself.  This was in the days of cabeese, no less.  The only reason the derailment was discovered as soon as it was was the fact that a signalling junction box was damaged, preventing the remote operation of a switch.

Granted, a derailment detector might have sensed the derailment and stopped the train, if the car was so equipped.  

On the other hand, the line would have been tied up until the car was manually re-railed.  As it was, the train arrived on time, and if it weren't for the damaged signal equipment, MOW would have been the first to discovere the evidence, perhaps days later...

[zugmann]

The beauty of railcars is their simplicity. It doesn't matter if you have a multi-billion dollar high-tech class 1 shop, or a weed-choked siding on uncle billy bob's shortline - both places can do just about any repairs needed on a railcar.  As was mentioned earlier, railcars are expected to perform no matter the temperature, humidity, weather, or experience of the crew slamming them around. 

 

When cars fall outside of the basic and proven form, their success has been limited. (Think roadrailers). Maybe if enough money and trials are given, there could be success.  But how much are we willing to invest?

 

 

Archimedes said "Give me a lever long enough and a fulcrum on which to place it, and I shall move the world." Just how long enough of a lever are we going to need?

[Euclid]

I few pages back, Paul North linked to information on derailment sensors that are entirely mechanical-pneumatic.  They do not require any data transmission.  However, one drawback that I see is that when one of these sensors sees a derailment, it dumps the air and dynamites the brakes.  As I mentioned, I believe this response to a derailment could actually cause a pileup that would not have happened without derailment sensors.   

One advantage that I see with ECP brakes is that they are not restricted to just an emergency brake application upon a derailment, or anything else that would cause the so-called UDE application.  Not only do ECP brakes have less potential for producing UDEs, but they also could be programmed to make a modulated application in response to what would normally cause a UDE, as opposed to making an emergency application.  ECP also has the advantage of applying all brakes simultaneously, as opposed to propagating the application in two directions from a mid-train UDE.

But with conventional air brakes, the full emergency application is all that is available in response to a derailment, or in response to a derailment sensed by a mechanical-pneumatic derailment detector.        

[Deggesty]

tree68

 

ruderunner

As for how an engineer could sense a derailment well isn't that what derailment sensors are for? Or perhaps its time to bring back rear end crews?  I know the snowball is getting bigger...

 

 

Reference the story in trains some time back (and a more recent incident I know of but won't get into), where a derailed car ran for several miles and then re-railed itself.  This was in the days of cabeese, no less.  The only reason the derailment was discovered as soon as it was was the fact that a signalling junction box was damaged, preventing the remote operation of a switch.

Granted, a derailment detector might have sensed the derailment and stopped the train, if the car was so equipped.  

On the other hand, the line would have been tied up until the car was manually re-railed.  As it was, the train arrived on time, and if it weren't for the damaged signal equipment, MOW would have been the first to discovere the evidence, perhaps days later...

 

Larry, I wish I could remember more details of something I observed in the summer of 1964--while going from Memphis to Chattanooga on Southern's #36, I stood at the rear while we traversed the area in which the railroad wanders back and forth between Tennessee and Georgia, just to note the state line signs, and noticed that many ties were gouged across the middle; the flagman told me that cars on one of the Southern freights had done the damage and, as I recall, he said that the errant cars did get back where they belonged.

[Euclid]

tree68

So, how fast was the Lynchburg train going when the derailment occurred? (It's probably been stated, but to review...)

 

Euclid

Therefore, if the trailing cars were to shove against the derailing cars, where would the resistance strong enough to cause jackknifing come from? 

 

 

That force would come from the first car that jacknifed going from X MPH to Zero MPH in about two seconds flat (or less).  The reason for the jacknife in the first place would be because the car dug into the ballast.  The faster it's going, the harder it's going to dig in.

The head end stopped when it did because the coefficient of friction between the cars and the ballast was much higher than what the brakes would have provided.  

1200 feet is less than a quarter mile - peanuts when you're talking about a loaded, 100 car (more or less) train.  

Once again - go back and look at the tornado derailment.  About six cars were blown off the track, amounting to some 300 feet.  I would opine that it's unlikely that even if there were 1200 feet between the locomotive and the cars still on the rail at the point that the derailed cars were clear of the rail, that the cars still on the rail would have stopped in time.  A derailment detector would have done nothing in that incident.

The remaining cars would clearly have been in emergency (or at least beginning the application), yet look at the speed with which the trailing cars collided with the locomotive.

Had that been an oil train, there would likely have been spills and a fire.

 

I think the ninth car did dig into the ballast.  You can see the track getting progressively torn up as the eight cars were dragged over it.  The ninth car derailed and jackknifed either before or after it separated from the eighth car.  If it jackknifed before it separated because the ninth car dug into the ballast, then that ninth car would have become the sole obstacle of resistance to cause the jackknife.  So the jackknifing would have begun between the ninth and tenth car as the ninth car dug in and became the obstacle to the tenth car.  That would have induced the eighth car to separate from the ninth car. 

Or the jackknifing may have come after the separation as the ninth car dug in and thus caused the separation prior to any jackknifing.

So the ninth car either began to jackknife and then separated because of that; or it separated first and jackknifed because it was no longer being pulled.  Either way, the jackknifing and the separation were nearly simultaneous, and either way, it does not change the point I am making about the derailment.  My point is that eight cars were pulled ahead and kept in line, and prevented from piling up.

I do not follow the point you make in comparing the tornado derailment to this Lynchburg derailment.  You say that a derailment sensor would not have helped.  I agree because the brakes went into emergency the instant the train derailed, so there would have been no advantage offered by quicker brake response as is the normal advantage of derailment sensors.    

But I do see two points in the tornado wreck that I had not considered before.  One is that the video shows the effect of derailed cars being dragged at high speed while staying coupled together and in line without digging in and jackknifing (just like with the Lynchburg wreck).  And those cars are being dragged while on their sides and completely off of the track bed.  That supports my ideas about the effect of differential braking just like the Lynchburg derailment does.

The other point I see in the tornado derailment is that the tank car is derailed and being shoved down the track without jackknifing.  I had considered that to be impossible, so I will have to revise my thinking about that type of action.  Granted, the track was apparently undamaged ahead of the tank car, but still I would not have expected a derailed car leading a cut to remain in line.  I suppose the derailed truck could have sat down on the rails in a way that continued the guidance by skidding on the rails. 

[zugmann]

Euclid

One advantage that I see with ECP brakes is that they are not restricted to just an emergency brake application upon a derailment, or anything else that would cause the so-called UDE application. Not only do ECP brakes have less potential for producing UDEs, but they also could be programmed to make a modulated application in response to what would normally cause a UDE, as opposed to making an emergency application. ECP also has the advantage of applying all brakes simultaneously, as opposed to propagating the application in two directions from a mid-train UDE.

 

Here's the million dollar question:  When do you program (and how?) the brakes to go into emergency vs. not?

For every one time where you may not want the brakes to dump, there's probably many more times when you do.

[BaltACD]

Euclid

 

tree68

So, how fast was the Lynchburg train going when the derailment occurred? (It's probably been stated, but to review...)

 

Euclid

Therefore, if the trailing cars were to shove against the derailing cars, where would the resistance strong enough to cause jackknifing come from? 

 

 

That force would come from the first car that jacknifed going from X MPH to Zero MPH in about two seconds flat (or less).  The reason for the jacknife in the first place would be because the car dug into the ballast.  The faster it's going, the harder it's going to dig in.

The head end stopped when it did because the coefficient of friction between the cars and the ballast was much higher than what the brakes would have provided.  

1200 feet is less than a quarter mile - peanuts when you're talking about a loaded, 100 car (more or less) train.  

Once again - go back and look at the tornado derailment.  About six cars were blown off the track, amounting to some 300 feet.  I would opine that it's unlikely that even if there were 1200 feet between the locomotive and the cars still on the rail at the point that the derailed cars were clear of the rail, that the cars still on the rail would have stopped in time.  A derailment detector would have done nothing in that incident.

The remaining cars would clearly have been in emergency (or at least beginning the application), yet look at the speed with which the trailing cars collided with the locomotive.

Had that been an oil train, there would likely have been spills and a fire.

 

 

 

I think the ninth car did dig into the ballast.  You can see the track getting progressively torn up as the eight cars were dragged over it.  The ninth car derailed and jackknifed either before or after it separated from the eighth car.  If it jackknifed before it separated because the ninth car dug into the ballast, then that ninth car would have become the sole obstacle of resistance to cause the jackknife.  So the jackknifing would have begun between the ninth and tenth car as the ninth car dug in and became the obstacle to the tenth car.  That would have induced the eighth car to separate from the ninth car. 

Or the jackknifing may have come after the separation as the ninth car dug in and thus caused the separation prior to any jackknifing.

So the ninth car either began to jackknife and then separated because of that; or it separated first and jackknifed because it was no longer being pulled.  Either way, the jackknifing and the separation were nearly simultaneous, and either way, it does not change the point I am making about the derailment.  My point is that eight cars were pulled ahead and kept in line, and prevented from piling up.

I do not follow the point you make in comparing the tornado derailment to this Lynchburg derailment.  You say that a derailment sensor would not have helped.  I agree because the brakes went into emergency the instant the train derailed, so there would have been no advantage offered by quicker brake response as is the normal advantage of derailment sensors.    

But I do see two points in the tornado wreck that I had not considered before.  One is that the video shows the effect of derailed cars being dragged at high speed while staying coupled together and in line without digging in and jackknifing (just like with the Lynchburg wreck).  And those cars are being dragged while on their sides and completely off of the track bed.  That supports my ideas about the effect of differential braking just like the Lynchburg derailment does.

The other point I see in the tornado derailment is that the tank car is derailed and being shoved down the track without jackknifing.  I had considered that to be impossible, so I will have to revise my thinking about that type of action.  Granted, the track was apparently undamaged ahead of the tank car, but still I would not have expected a derailed car leading a cut to remain in line.  I suppose the derailed truck could have sat down on the rails in a way that continued the guidance by skidding on the rails. 

 

You cannot calculate what will happen in a derailment BEFOREHAND.  You can see what happened afterwards, but the instant to instant of what is happending during the incident cannot be accurately predicted.

Just like ACCURATELY predicting earthquakes; you can calculate the stresses afterwards, but have no REAL handle on the stresses beforehand and what the actual breaking point is and where the weakest point is.

[Euclid]

zugmann

 

Euclid

One advantage that I see with ECP brakes is that they are not restricted to just an emergency brake application upon a derailment, or anything else that would cause the so-called UDE application. Not only do ECP brakes have less potential for producing UDEs, but they also could be programmed to make a modulated application in response to what would normally cause a UDE, as opposed to making an emergency application. ECP also has the advantage of applying all brakes simultaneously, as opposed to propagating the application in two directions from a mid-train UDE.

 

 

 

Here's the million dollar question:  When do you program (and how?) the brakes to go into emergency vs. not?

For every one time where you may not want the brakes to dump, there's probably many more times when you do.

 

I would say they should go into a non-emergency braking mode any time there is an application initiated by the system without the intent of the engineer.  Then the engineer would have the ability to override that application with maximum braking if it were deemed necessary.  I am speaking of ECP brakes without any differential braking system that I have descibed.  If you had differential braking, the ECP would automatically go into that mode for all unintended brake applications, and that too could be overriden if necessary. 

[NorthWest]

I may be misunderstanding what you are saying, but how can the engineer tell how cars are derailing in the middle of his train? It would likely be out of sight.

[Euclid]

NorthWest

I may be misunderstanding what you are saying, but how can the engineer tell how cars are derailing in the middle of his train? It would likely be out of sight.

The engineer would not see the derailment happening in many cases, and I don’t mean to suggest that he do anything with the braking in response to the derailment.  The system with ECP and derailment sensors would detect a derailment and automatically react with a default brake application.  That would be something less than a full force application at first, and then ramp up to full force.  The point would be to perturb the derailment as little as possible with the brake application so as not to encourage the development of jackknifing and a pileup.  I would call that a modulated brake response.

If the system were further equipped with differential braking, upon sensing a derailment, the default brake response would be to apply more brake force to the cars behind the derailment than to the cars ahead of it.  The point would be to prolong the car dragging for as long as possible without a jackknifing and pileup.  The longer the dragging, the more kinetic energy can be dissipated before a pileup begins.  So it works to either prevent a pileup, or to reduce the energy that is fed into a pileup if one occurs.

With either the modulated braking or differential braking, the train stopping distance is increased because full braking force is withheld to some extent.  So, just in case the train derails and starts this specialized braking, and the engineer happens to spot a gasoline tanker truck stalled on a grade crossing 2000 feet ahead; he can override the automatic default ECP derailment braking to stop before hitting the gasoline truck.

That is the reason for the engineer having the option to override the automatic default application caused by a derailment. The default application would be only in response to a derailment, and to any application not intended by the engineer.     

 

[ruderunner]

tree68

 

 

ruderunner

As for how an engineer could sense a derailment well isn't that what derailment sensors are for? Or perhaps its time to bring back rear end crews?  I know the snowball is getting bigger...

 

 

Reference the story in trains some time back (and a more recent incident I know of but won't get into), where a derailed car ran for several miles and then re-railed itself.  This was in the days of cabeese, no less.  The only reason the derailment was discovered as soon as it was was the fact that a signalling junction box was damaged, preventing the remote operation of a switch.

Granted, a derailment detector might have sensed the derailment and stopped the train, if the car was so equipped.  

On the other hand, the line would have been tied up until the car was manually re-railed.  As it was, the train arrived on time, and if it weren't for the damaged signal equipment, MOW would have been the first to discovere the evidence, perhaps days later...

 

Larry that sounds like just dumb luck. And frankly the time lost manually rerailing the car pales in comparison   to the cost of wreck cleanup and repair of the track if the derailment got worse.

 

Here's a thought for a cheap simple and rugged derailment sensor: a set of limit switches located near the kingpin and contacted by an arm mounted to the truck. Set them so that the arm contacts a switch if the truck rotates too far from straight ahead. Example if one truck rotates 10 degrees but the other stays straight ahead that indicates a truck off the rails. 10 degrees is just an example number, the actual number would vary by car length and track curvature. This wouldn't have to effect a brake application but just an alarm to alert the crew. Heck it could even include a "switching" setting to prevent false alarms when working tight trackage at low speeds. It doesn't need to be extremely precise either, if the truck is within normal range of rotation then its likely tracking correctly. But once outside of that range we need to stop the train and find out what happened.

I'm not looking to prevent derailments but rather prevent simple wheels off rails from turning into a flaming pileup.

[ruderunner]

Zug maybe railroaders need to adapt to changing technology. Maybe they need to learn a new skill set. And if they are already smart enough the repair a locomotive then this newer technology shouldn't completely stymie them. Geez look at the computer controls on an ac44 for example. That makes the rest of what were discussing nearly stone age in terms of complexity.

I kind of understand the resistance to change and technology, I deal with it daily. Then I hop into my non computerized car and drive home. I dont care for vehicles that coddle the driver and activate brakes and steering for them. But those drivers would be just as well served to ride the bus,er, tram.

Let's face it though, train brakes haven't changed too much since old man Westinghouse came up with air brakes 100+years ago. Its time to take another look.

[dehusman]

Euclid

That would be something less than a full force application at first, and then ramp up to full force.  The point would be to perturb the derailment as little as possible with the brake application so as not to encourage the development of jackknifing and a pileup.  I would call that a modulated brake response.

You just don't understand the train dynamics.  The emergency application isn't "perturbing" the derailment, the derailed cars are perturbing the derailment.  You keep forgetting gravity.  its gravity that you are fighting.  All the pulling power in the world won't overcome gravity.  The derailed cars are going to go down into the roadbed, if the cars get off the track structure the cars are going down into the ditches.  Gravity.  You can pull all you want and its not going to help a bit. 

 

Its not that the 8 cars were magically pulled to safety, its that it took 8 cars dragging over the POD to get a big enough hole to cause the jacknifing.

[BaltACD]

Euclid & Rude - sell the FRA on your ideas.  Simple!  Once they are sold it can be mandated by the end of the year!

[cx500]

Don't you see your example of the computer controls on the AC4400 (or equivalent EMD for that matter) shows how the railroaders do adapt, and enthusiastically embrace new technology.  Consider DPU operation; it has become widespread, but it took years of development and technological refinements before it became fully reliable.  What once required bulky equipment in a dedicated piece of rolling stock is now merely a piece of software in a micro-processor.  Those track geometry cars that can measure track surface, gauge, rail wear while travelling at 60mph are another example of how, in this case, the track department is embracing new technology as soon as it becomes available.

The railroads are not resistant to change, or technology.  They do, however, want to ensure that any new development is robust enough to enhance safe operations in a way that is both close to 100% reliable and affordable.  If the benefit is marginal and the cost is significant there are better ways to improve safety.

Several railroads have tested electric controlled braking systems on dedicated unit trains.  They were willing to commit a fairly significant investment to outfit several hundred cars and a number of locomotives, and run them in regular service to gain experience with the benefits and problems that may occur in real world operation.  It would appear that the test results were not compelling.

Many things would be "nice to have" but cost is always an issue.  A few thousand dollars a car doesn't sound like much, until you multiply it by the number of cars roaming the rails of North America.  It just became billions of dollars of extra cost. 

While home computers have become very powerful and affordable, try using one outside in the pouring rain, or a dust storm, and see how long it remains trouble free.  Drop it a few times each day to simulate slack action and bouncing over frogs and diamonds.  Any electronics on freight equipment face a tough environment.  Ideally they should survive trouble free for the 50-year lifespan of the car.  And that's not realistic.

[Euclid]

Dave,

If you would just set aside your personal attacks for a couple minutes and read what I have said, maybe you would understand it better.  The way you are lumping all kinds of things together creates a tangle that is hard to respond to.  

You say things like, “All the pulling power in the world won't overcome gravity.”  Where do I begin to respond to that?

You say, “Its not that the 8 cars were magically pulled to safety, its that it took 8 cars dragging over the POD to get a big enough hole to cause the jacknifing.”

We have talked about that and I am not sure why you think it refutes what I am saying.  I never said that I expect every attempt at pulling the train to prevent every bit of jackknifing.

You say, “Pouring more kinetic energy into a wreck is not helpful. The idea is to get everything stopped as fast as possible. If derailment detectors do that, that contributes. If ECP does that, then that contributes. Digging a bigger hole in the track faster doesn't help anything.”

I have not suggested pouring more kinetic energy into a wreck.  I never suggested solving the problem by digging a bigger hole in the track. 

I would say that preventing tank car breaching is more important than getting everything stopped as fast as possible.  The eight cars that did not jackknife in the Lynchburg wreck could just as easily have jackknifed, and added fuel to the fire.  If they had, the train probably would have stopped quicker than it did.  How would that have been a benefit? 

[ruderunner]

cx500

Don't you see your example of the computer controls on the AC4400 (or equivalent EMD for that matter) shows how the railroaders do adapt, and enthusiastically embrace new technology.  Consider DPU operation; it has become widespread, but it took years of development and technological refinements before it became fully reliable.  What once required bulky equipment in a dedicated piece of rolling stock is now merely a piece of software in a micro-processor.  Those track geometry cars that can measure track surface, gauge, rail wear while travelling at 60mph are another example of how, in this case, the track department is embracing new technology as soon as it becomes available.

The railroads are not resistant to change, or technology.  They do, however, want to ensure that any new development is robust enough to enhance safe operations in a way that is both close to 100% reliable and affordable.  If the benefit is marginal and the cost is significant there are better ways to improve safety.

Several railroads have tested electric controlled braking systems on dedicated unit trains.  They were willing to commit a fairly significant investment to outfit several hundred cars and a number of locomotives, and run them in regular service to gain experience with the benefits and problems that may occur in real world operation.  It would appear that the test results were not compelling.

Many things would be "nice to have" but cost is always an issue.  A few thousand dollars a car doesn't sound like much, until you multiply it by the number of cars roaming the rails of North America.  It just became billions of dollars of extra cost. 

While home computers have become very powerful and affordable, try using one outside in the pouring rain, or a dust storm, and see how long it remains trouble free.  Drop it a few times each day to simulate slack action and bouncing over frogs and diamonds.  Any electronics on freight equipment face a tough environment.  Ideally they should survive trouble free for the 50-year lifespan of the car.  And that's not realistic.

 

cx500. That's kind of my point. This technology isn't a great leap forward, its not much different than what already exists.

Also I'm not really pushing for ECP .

Nor did I say a home PC would be applicable to railcar use, it was just an example of how technology costs come down as useage goes up. You yourself pointed out that former DPU controls used to require a separate CSR to contain them and now fit comfortably into the operator's console.

I think there is a disconnect between what Euclid is attempting to describe and  how everyone is perceiving it. He's not advocating adding kinetic energy but rather controllinig it, and attempting to dissapate it in directions that will help minimize how much goes into a pileup.

 

[oltmannd]

ruderunner

Here's a thought for a cheap simple and rugged derailment sensor: a set of limit switches located near the kingpin and contacted by an arm mounted to the truck. Set them so that the arm contacts a switch if the truck rotates too far from straight ahead. Example if one truck rotates 10 degrees but the other stays straight ahead that indicates a truck off the rails. 10 degrees is just an example number, the actual number would vary by car length and track curvature. This wouldn't have to effect a brake application but just an alarm to alert the crew. Heck it could even include a "switching" setting to prevent false alarms when working tight trackage at low speeds. It doesn't need to be extremely precise either, if the truck is within normal range of rotation then its likely tracking correctly. But once outside of that range we need to stop the train and find out what happened. I'm not looking to prevent derailments but rather prevent simple wheels off rails from turning into a flaming pileup.

No doubt something like this could be worked out. I'd look for rate of change rather than absolute limits.  It could react faster.

Lots of research needed, however.  Not simple or easy to do.  Expensive to validate since you have to actually derail a car at speed, on purpose.

But, that's not the biggest problem.  The biggest problem is failed compontents.  Mainline derailments are exceedingly rare.  Component failures are not.  How you mount them, wire them, etc.  Are you allowed to run with failed components?  What do you do when a train stops for a failed sensor?

The trick is not to create more trouble than you solve.