Longer Trains Cause More Derailments

[Euclid]

blue streak 1

Instead of blaming longer trains let us measure derailments by car miles.  What is the derailment numbes by say million car miles?  Each derailment can go into a data base for its length.  Switching moves are probably the easiest types to define.

 

The question I asked in the title of this thread does reflect a growing concern that the relative large increase in average freight train length is causing an increase in derailments.  The prospect has not been sufficiently investigated to determine whether the cause is train length in number of feet, or train length in number of cars.  Probably both play a role.

 

The presumed explanation is that the increased length of freight trains tends to cause higher in-train forces related to braking and power changes combined with slack action.  So the higher forces are primarily related to slack run-in and run-out causing spikes in the forces of buff (compression) and draft (tensile).  Too much buff force causes a train to buckle and jackknife.  Too much draft force “stringlines” a train on curves or pulls a train in two on straight track.   

 

Another possibly contributing factor is the preferred use of dynamic braking instead of using the automatic air brake.  Dynamic braking only applies brakes on the locomotive of a train.  In the days of 100-car trains with no DPU; and only head end engines in M.U.; dynamic braking was only a secondary type of braking, with the primary braking being done with the automatic air brakes. 

 

The reason was that automatic air braking spread the braking throughout the 100 cars of the train.  Whereas dynamic braking turned the multiple locomotives, which were only on the head end, into a giant anchor with the entire consist of 100 fee rolling cars shoving up against that anchor.  So unless the dynamic braking application was extremely limited, it would cause a significant risk of a derailment due to high buff force relatively nearer to the head end.  And if the dynamic braking was sufficiently limited to avoid such a derailment, it was simply not enough braking to safely control the train. 

 

However, in this era with DPU, that problem of excess buff force from dynamic braking has been greatly diminished because the DPU also distributes the dynamic braking.  But this distribution is still not thorough like it is with automatic air braking applying to each car in the train. 

 

Regarding the question of the number of derailments per car mile, the answer might be useful, but it is an entirely different issue, which is related to railcar construction.  It would not be an alternative to the question of the increasing number of derailments caused by growing train length raising in-train forces. 

 

The question that I asked in the original title is being researched for an answer by the FRA.  This cannot be easily done with test trains run to a point of derailment.  It has to be done with test trains run with sensors to record in-train forces.  The data from this could be used in the development of new computer test programs.  Then their results can be double-checked in more test trains.  Eventually, they will have a program that will accurately predict the in-train forces of any given train makekup, in any length of train. Then they will know what the limits are.

[diningcar]

To answer this: first define a longer train - 2000-4000-8000- 12000 feet. It becomes subjective and subject to biases which make it Imposible to satisfy all. 

[Overmod]

NittanyLion

They don't expect or demand that level of experience.  Day one experience should do it, but even that failed.  Air France 447 was lost because the pilot didn't know it was possible to stall an A330.  And the guy in the cockpit with the most experience never bothered to cross-check the physical controls until it was well past the point of no return. 

That almost directly proves Zug's point, though.  No experienced pilot would ever say 'you can't stall an A330' -- the problem was that the person didn't think the aircraft was stalled when the fancy computers were telling him it wasn't.

And there's something well-established in cognitive science that's worse: when you have to constantly watch and second-guess an automatic control system, which may not be explaining its 'decisions', waiting hours and hours against the chance there might suddenly be 30 seconds of screaming terror.  (To say nothing about doing it on little sleep, outside circadian rhythm, after a night of sleep apnea with CPAP...)

That goes a long way toward establishing he's right when 'you can't have it both ways' -- just has been the case with automated highway design since the late Forties.

[NittanyLion]

zugmann

But when those systems crap out, or something huge and out of the ordinary pops up, then the company expects the pilots (or engineers) to have the skills of a 30-yr man that has been running every day.   Can't. Have. It. Both. Ways.  

They don't expect or demand that level of experience.  Day one experience should do it, but even that failed.  Air France 447 was lost because the pilot didn't know it was possible to stall an A330.  And the guy in the cockpit with the most experience never bothered to cross-check the physical controls until it was well past the point of no return.  

[BaltACD]

blue streak 1

Instead of blaming longer trains let us measure derailments by car miles.  What is the derailment numbes by say million car miles?  Each derailment can go into a data base for its length.  Switching moves are probably the easiest types to define.

Somewhere along the line the FRA keeps such statistics.

[blue streak 1]

Instead of blaming longer trains let us measure derailments by car miles.  What is the derailment numbes by say million car miles?  Each derailment can go into a data base for its length.  Switching moves are probably the easiest types to define.

[Overmod]

.

Reply was meant to go to a different thread.

[Euclid]

Cotton Belt MP104

As I went through this thread there were remarks that seemed to indicate some unkind remarks had been made. I was subject to that at one time and refused to post anything.

As I have said before if a participant is seemingly pretending to be an expert and is not, the method of dealing with the bulk of those comments can be ignored (I mentioned "extinction"). While it is true that the "dull and ignorant, they too have their story and should be heard to an extent. This is only civil behavior. But when the story and details are of extraordinary length one seems to not gain that much, even a flaw in information can be present (intentional or not) and that draws a reaction from those who know better....and the verbal fight is on.

Example: Posted by Euclid on Thursday, June 1, 2023 10:02 AM : In the comment there is a statement:   ““If you use air brakes, then all 840 car wheels of the train will have brake shoes pressing into them.”  Is that a not true statement?”  

 

Cotton Belt MP104: 

 

Even with the FRED which sends the “signal to brake” from rear as front is sending backward the signal to brake. This is a cascading operation and not immediate “all 840 wheels…have brake shoes pressing” ?  He Did Ask The Question is that not a true statement?  And I was wondering if I was right that the statement is not right.  Endmrw0621231426

 

I answered your question (highlighted in red), but you never responded, so just to clarify:

 

You were not right in your statement that seems to say you thought my statement about the number of brake shoes applying was not right.

 

You have quoted my original statement without its clarifying context which was, “It is only to distinguish the effects of air brakes compared to dynamic brakes.”

 

The point is the difference in applying air brakes which applies brake shoes on all 210 cars to slow them, versus applying only dynamic brakes that retard the wheels of only the leading two engines to slow 210 cars. 

[rdamon]

zugmann

 

 

rdamon

Talking to Airbus drivers they have the equivilent of Trip Optimizer. Climb, Cruise and Decent is all managed by the FMS (Flight Management System) computer. Do it yourself and get out line, expect a call from your union rep as the system will phone home. 

 

 

But when those systems crap out, or something huge and out of the ordinary pops up, then the company expects the pilots (or engineers) to have the skills of a 30-yr man that has been running every day.  

Can't. Have. It. Both. Ways.  

 

Agreed, but it will probably take a major incident to change behavior.  Look at the 737-MAX

[Cotton Belt MP104]

 

 Zug.   Good info. My reference to specific flight (my bad) was only using what was mentioned earlier. Sine you have the video of the girl being run over, you probably have the correct flight info. I will be looking up BOTH incidents. Famous hero Schullenberger (sp?) sure did his best when things didn't go right. To keep this on RR topic...same can be said of RR engineers who operate by the "seat of his pants" in otherwords (gauges important also) what he feels that can't be measured makes a HUGE difference in his moves.

interesting detail about infared detector and rules not obeyed endmrw0623231449

your comments below

Asiana Flight 214?  3 killed, 180-some injured. You may be thinking of another incident? 

And they knew the one girl was laying in the grass before they sprayed foam (video proof - easily avaliable).  Then they sprayed foam and ran her over - made all the worse because the truck didn't have its thermo cameras installed yet, depsite the FAA regulation on same. 

 

[/quote]

[BaltACD]

Euclid

... 

But for autonomous operation there is no train crew.  Without a train crew, there is no reason to run monster trains in order to move more cars with one crew.  If anything, autonomous running implies shorter trains of say 50-75 cars maximum to take advantage of all the virtues of short trains with their more nimble operation.  Maybe then we could have an operation actually looks like “Precision Scheduled Railroading.”

Wrong again!  Track capacity is track capacity no matter if the trains are crewed or not.  It is exceedingly easy to unleash too many trains on a particular track segement to permit fluid operations.  

How do these autonomous trains perform the 'block swappng' that is one of the hallmarks of the PSR operating plan - setting off and picking up on line of road?  Even in pre PSR operating plans trains pick up and set off at points along their runs.

[Euclid]

Cotton Belt MP104

As I went through this thread there were remarks that seemed to indicate some unkind remarks had been made. I was subject to that at one time and refused to post anything.

As I have said before if a participant is seemingly pretending to be an expert and is not, the method of dealing with the bulk of those comments can be ignored (I mentioned "extinction"). While it is true that the "dull and ignorant, they too have their story and should be heard to an extent. This is only civil behavior. But when the story and details are of extraordinary length one seems to not gain that much, even a flaw in information can be present (intentional or not) and that draws a reaction from those who know better....and the verbal fight is on.

Example: Posted by Euclid on Thursday, June 1, 2023 10:02 AM : In the comment there is a statement:   ““If you use air brakes, then all 840 car wheels of the train will have brake shoes pressing into them.”  Is that a not true statement?”  

 

Even with the FRED which sends the “signal to brake” from rear as front is sending backward the signal to brake. This is a cascading operation and not immediate “all 840 wheels…have brake shoes pressing” ?  He Did Ask The Question is that not a true statement?  And I was wondering if I was right that the statement is not right.  Endmrw0621231426

 

The topic of the thread title simply asks whether the recent trend of large increases of train length made possible by distributed power are having the unintended consequence of causing an increase in derailments.  This possibility is based on the fact that the greater train length increases the amount of train slack action, and that leads to greater in-train forces related to slack running in and out. 

 

The FRA believes that the answer to the question is:  “Yes the longer trains have a higher probability of derailing.”  They have released a report on this matter.  In the report, they also cite three recent derailments which they review in detail, and conclude that they were caused by excess in-train forces.

 

One of them is the Norfolk Southern derailment that occurred at Springfield, OH, a couple weeks after the East Palestine derailment.  I have posted the FRA analysis of this wreck several times in this thread, but so far nobody has indicated that they have read it.

 

Essentially, the FRA says that the 210-car train had two engines on-line on the head end, which were being used for dynamic braking.  There was also a DPU engine mid-train, but it was not being used for dynamic braking.  The FRA report leaves the impression that this head end dynamic braking by two engines is the only braking being used at the time. 

 

FRA also says that the train weight was relatively higher toward the front and rear of the train; and was lower near the middle of the train. 

 

At the time of the derailment, the train was moving around 45 mph, and was approximately centered on a sag in the track profile.  So the front 105 cars of the train was running uphill, and the rear 105 cars were running downhill while being retarded by dynamic braking.  The engineer was increasing the dynamic brakes to slow the train while passing through the town.  Incidentally, I would like to know whether the FRA is saying that this dynamic braking was the only braking in effect at the time.  They imply that was the case.  If so, you have a 210-car train being slowed only by increasing the turning resistance of the wheels of two locomotives on the head end of the train.

 

At the moment of derailment, the front half of the train was running uphill with slack compression building rearward thought its cars as dynamic braking propagated back from the head end.  Also adding retardation to the front half of the train was the fact that it was running uphill and thus being slowed by gravity in addition to dynamic braking.

 

Meanwhile, the bunching slack in the uphill running head end had not yet reached the second half of the train which was freely rolling downhill and accelerating under the force of gravity without any braking, and with slack generally loose.  So, within  the range of slack, the uphill head end was decelerating and the downhill hind end was accelerating.  This is leading to a form of head-on collision that would be very interesting to study with computer modeling.   

 

As the moment of derailment approached, the head-end slack had begun to continue its running into the second half of the train.    

 

At that moment, two empty coil steel cars entered a grade crossing at the bottom of the sag.   At that same moment, the buff force of compression advancing from front to rear, reached its highest pressure, and caused the train to buckle and derail between the two steel coil cars.

 

+++++++++++++++++++++++++++++++

 

 

My point about the number of brake shoes touching the car wheels was only to illustrate that with dynamic brakes being used, there are no brake shoes touching the wheels.  I even included that clarification along with stating the point as posted further up on this page.   I made this point specifically in reference to the Springfield, OH wreck that I reviewed above.  Using only dynamic brakes from only two head-end locomotives, as the sole braking for a 210-car train with sub-optimal train make-up; in territory with “hogback” or undulating track profile--- This combination strikes me as being extremely risky.   

 

To the question of whether human operator skill is better than computer autonomous operation, that point has not much come up with this topic of longer trains causing more derailments due to excess in-train forces.  However, a piece of the in-train force puzzle is train make-up, and computer programs have been developed to make the decisions for optimal placement into the train consist, of various types and weights of railcars.  Like all programs, all they need is more work. 

 

But for autonomous operation there is no train crew.  Without a train crew, there is no reason to run monster trains in order to move more cars with one crew.  If anything, autonomous running implies shorter trains of say 50-75 cars maximum to take advantage of all the virtues of short trains with their more nimble operation.  Maybe then we could have an operation actually looks like “Precision Scheduled Railroading.”

[zugmann]

rdamon

Talking to Airbus drivers they have the equivilent of Trip Optimizer. Climb, Cruise and Decent is all managed by the FMS (Flight Management System) computer. Do it yourself and get out line, expect a call from your union rep as the system will phone home. 

But when those systems crap out, or something huge and out of the ordinary pops up, then the company expects the pilots (or engineers) to have the skills of a 30-yr man that has been running every day.  

Can't. Have. It. Both. Ways.  

[zugmann]

Cotton Belt MP104

Sad but true. The plane "crashed" but foam was on the runway to prevent fire. No one was injured, EXCEPT one young lady perished as she was in the foam and run over by a firetruck. endmrw0622232119

Asiana Flight 214?  3 killed, 180-some injured. You may be thinking of another incident? 

And they knew the one girl was laying in the grass before they sprayed foam (video proof - easily avaliable).  Then they sprayed foam and ran her over - made all the worse because the truck didn't have its thermo cameras installed yet, depsite the FAA regulation on same. 

[rdamon]

Backshop

 

 

The same thing happened with the Asiana flight into San Francisco.  My brother is a retired Delta A320 captain and was often frustrated by some of the younger pilots. They didn't want to "fly" the airplane.  My brother started as an army chopper pilot and enjoyed handflying planes.

 

 

Talking to Airbus drivers they have the equivilent of Trip Optimizer. Climb, Cruise and Decent is all managed by the FMS (Flight Management System) computer. Do it yourself and get out line, expect a call from your union rep as the system will phone home. 

[Cotton Belt MP104]

Reference Asiana/flight/SF

Sad but true. The plane "crashed" but foam was on the runway to prevent fire. No one was injured, EXCEPT one young lady perished as she was in the foam and run over by a firetruck. endmrw0622232119

[Backshop]

Cotton Belt MP104

 

As an example: Airlines (look it up: Children of Magenta) have had pilots depend on auto pilot so much that they don't take off the auto pilot (when seat of the pants assesment of their situational awarness indicates, WE ARE IN TROUBLE.). They continue to depend on the automation with disasterous results. Some pilots have warned/commented, "For heaven sakes, you learned to fly a plane without auto pilot, TURN that darn computer off and fly it like you know a plane should be flown" endmrw0621231354 

 

The same thing happened with the Asiana flight into San Francisco.  My brother is a retired Delta A320 captain and was often frustrated by some of the younger pilots. They didn't want to "fly" the airplane.  My brother started as an army chopper pilot and enjoyed handflying planes.

[BaltACD]

Cotton Belt MP104

Not being an locomotive engineer, I can not speak from experience.

But I have read enought and logic would make sense that due to the many variable forces in a rolling train, operating by the "seat of your pants" (as the expression goes) is far superior to automation, while nice, automation has limits to variables that are not programed in, only the basics which might be several. BUT....Ain't nutin' compared to years of experience.

As an example: Airlines (look it up: Children of Magenta) have had pilots depend on auto pilot so much that they don't take off the auto pilot (when seat of the pants assesment of their situational awarness indicates, WE ARE IN TROUBLE.). They continue to depend on the automation with disasterous results. Some pilots have warned/commented, "For heaven sakes, you learned to fly a plane without auto pilot, TURN that darn computer off and fly it like you know a plane should be flown" endmrw0621231354 

For any that are interested - the Mentour Pilot YouTube channel has a series of accident investigation videos - Very enlightening.

https://www.youtube.com/watch?v=ylLjnLh_XA4&t=27s

 

 

Note - the plane crashed on the Trans-Siberian Railroad.

[Cotton Belt MP104]

As I went through this thread there were remarks that seemed to indicate some unkind remarks had been made. I was subject to that at one time and refused to post anything.

As I have said before if a participant is seemingly pretending to be an expert and is not, the method of dealing with the bulk of those comments can be ignored (I mentioned "extinction"). While it is true that the "dull and ignorant, they too have their story and should be heard to an extent. This is only civil behavior. But when the story and details are of extraordinary length one seems to not gain that much, even a flaw in information can be present (intentional or not) and that draws a reaction from those who know better....and the verbal fight is on.

Example: Posted by Euclid on Thursday, June 1, 2023 10:02 AM : In the comment there is a statement:   ““If you use air brakes, then all 840 car wheels of the train will have brake shoes pressing into them.”  Is that a not true statement?”  

 

Even with the FRED which sends the “signal to brake” from rear as front is sending backward the signal to brake. This is a cascading operation and not immediate “all 840 wheels…have brake shoes pressing” ?  He Did Ask The Question is that not a true statement?  And I was wondering if I was right that the statement is not right.  Endmrw0621231426

[Cotton Belt MP104]

Not being an locomotive engineer, I can not speak from experience.

But I have read enought and logic would make sense that due to the many variable forces in a rolling train, operating by the "seat of your pants" (as the expression goes) is far superior to automation, while nice, automation has limits to variables that are not programed in, only the basics which might be several. BUT....Ain't nutin' compared to years of experience.

As an example: Airlines (look it up: Children of Magenta) have had pilots depend on auto pilot so much that they don't take off the auto pilot (when seat of the pants assesment of their situational awarness indicates, WE ARE IN TROUBLE.). They continue to depend on the automation with disasterous results. Some pilots have warned/commented, "For heaven sakes, you learned to fly a plane without auto pilot, TURN that darn computer off and fly it like you know a plane should be flown" endmrw0621231354 

[BaltACD]

Don't forget the 'What about?'

[Overmod]

BaltACD

Psychot

For those of us who have zero experience in the railroad industry, it's invaluable to have railroaders in this forum sharing their experiences. Having said that, some of you might want to consider being a bit less condescending in your replies. Educate us, rather than simply typing the equivalent of "you don't know $hit."

There comes a point in time, after repeated attempts to educate a particular participant without success, one has to come to the realization that the participant 'don't know s..t' and doesn't want to know.

But the specific thing he's talking about is the comment from Big Jim, that was dismissive to him without explaining why.  All he's asking for is that the reply include why what he posted was 'wrong' -- even if that's just a couple of lines, an URL for the current Al Krug site, or whatever.

[charlie hebdo]

I think Sam's comments of first-hand experience with handling very long trains is definitive.  Longer trains (actually they are often several trains coupled together to reduce labor costs) are more prone to derailments.

[Euclid]

SFbrkmn

Many of these mega trains are not just bulk unit or intermodal movements, but mixed freight that is asking for a nightmare to happen. 25 yrs on the rails and these trains make me nervous. Crew has to carefully look over the profile tonnage to aboslutely make sure the train is in tonnage compliance. Some trains are barely in compliance. Either way, it is nothng but a five digit footage slingshot banging you around in the seat hoping you have armrests to prevent being tossed onto the floor making for a miserable trip w/the hope auto control does not create a break-in two of a knuckle, drawbar or derailment. Regardless of it is a auto or manual control, any breakdown generates the self asked thought "Should have laid off for this trip".

Sam 

 

 

Sam,

 

Your point is well made when you refer to monster trains as being, “nothing but a five digit footage slingshot banging you around in the seat hoping you have armrests to prevent being tossed onto the floor…”

 

The long slack action can allow a portion of the train to roll faster than the rest of it even though all the cars are coupled together.  So, in effect, in one long train, you can have two different, independent trains; and a collision between them.

 

The longer the whole train is, the larger those two independent trains are, and the more violent the collision can be.  So it follows that the potential for relatively longer trains to have a greater chance of derailing due to in-train forces is a fact.    

[BaltACD]

Psychot

For those of us who have zero experience in the railroad industry, it's invaluable to have railroaders in this forum sharing their experiences. Having said that, some of you might want to consider being a bit less condescending in your replies. Educate us, rather than simply typing the equivalent of "you don't know $hit."

There comes a point in time, after repeated attempts to educate a particular participant without success, one has to come to the realization that the participant 'don't know s..t' and doesn't want to know.

[Psychot]

For those of us who have zero experience in the railroad industry, it's invaluable to have railroaders in this forum sharing their experiences. Having said that, some of you might want to consider being a bit less condescending in your replies. Educate us, rather than simply typing the equivalent of "you don't know $hit."

[SFbrkmn]

Many of these mega trains are not just bulk unit or intermodal movements, but mixed freight that is asking for a nightmare to happen. 25 yrs on the rails and these trains make me nervous. Crew has to carefully look over the profile tonnage to aboslutely make sure the train is in tonnage compliance. Some trains are barely in compliance. Either way, it is nothng but a five digit footage slingshot banging you around in the seat hoping you have armrests to prevent being tossed onto the floor making for a miserable trip w/the hope auto control does not create a break-in two of a knuckle, drawbar or derailment. Regardless of it is a auto or manual control, any breakdown generates the self asked thought "Should have laid off for this trip".

Sam 

[Euclid]

tree68

One might conclude that while longer trains may be a factor in derailments, train make-up and handling may be the bigger piece.

Would the Springfield derailment have happened if everything behind the middle of the train was empties?  Rhetorical question, but a consideration.

 

I think that it is train makeup and handling that are the typical final triggers in derailing ultra-long trains.  That seems to be the position of the FRA as well.  But the FRA position also seems to be that it is the extreme train length that is the underlying cause.  Clearly they say that ultra-long trains are more difficult for engineers to control and operate without derailing, compared to the conventional size trains.

 

They also say that is not acceptable to simply allow engineers who are experienced only with conventional trains to learn to operate the ultra-long trains just by gaining on the job experience with them.  The FRA says the differences between the conventional trains and ultra-long trains is so great that engineers must receive new classroom instruction and training before being allowed to operate them on the job. 

 

To your rhetorical question:  I think it is quite possible or likely that the derailment would not have happened if the trailing half of the consist was all empties.  Also the derailment was far less likely had the trailing half not been descending the sag at the same time the leading half was ascending the sag.  And also I think there is a good chance that the derailment would not have happened if dynamic braking were being applied from both the head end and from the mid-train DPU, rather just all from the head end.

[tree68]

One might conclude that while longer trains may be a factor in derailments, train make-up and handling may be the bigger piece.

Would the Springfield derailment have happened if everything behind the middle of the train was empties?  Rhetorical question, but a consideration.